fix scaling 1

analyze_memory_profile.py

@@ -0,0 +1,193 @@
+#!/usr/bin/env python3
+"""
+Analyze memory profile JSON files to identify OOM causes
+"""
+
+import json
+import glob
+import os
+import sys
+from pathlib import Path
+
+def analyze_memory_files():
+    """Analyze partial memory profile files"""
+    
+    # Get all partial files in order
+    files = sorted(glob.glob('memory_profile_partial_*.json'))
+    
+    if not files:
+        print("❌ No memory profile files found!")
+        print("Expected files like: memory_profile_partial_0.json")
+        return
+        
+    print(f"🔍 Found {len(files)} memory profile files")
+    print("=" * 60)
+    
+    peak_memory = 0
+    peak_vram = 0
+    critical_points = []
+    all_checkpoints = []
+    
+    for i, file in enumerate(files):
+        try:
+            with open(file, 'r') as f:
+                data = json.load(f)
+            
+            timeline = data.get('timeline', [])
+            if not timeline:
+                continue
+                
+            # Find peaks in this file
+            file_peak_rss = max([d['rss_gb'] for d in timeline])
+            file_peak_vram = max([d['vram_gb'] for d in timeline])
+            
+            if file_peak_rss > peak_memory:
+                peak_memory = file_peak_rss
+            if file_peak_vram > peak_vram:
+                peak_vram = file_peak_vram
+                
+            # Find memory growth spikes (>3GB increase)
+            for j in range(1, len(timeline)):
+                prev_rss = timeline[j-1]['rss_gb']
+                curr_rss = timeline[j]['rss_gb']
+                growth = curr_rss - prev_rss
+                
+                if growth > 3.0:  # >3GB growth spike
+                    checkpoint = timeline[j].get('checkpoint', f'sample_{j}')
+                    critical_points.append({
+                        'file': file,
+                        'file_index': i,
+                        'sample': j,
+                        'timestamp': timeline[j]['timestamp'],
+                        'rss_gb': curr_rss,
+                        'vram_gb': timeline[j]['vram_gb'],
+                        'growth_gb': growth,
+                        'checkpoint': checkpoint
+                    })
+                    
+            # Collect all checkpoints
+            checkpoints = [d for d in timeline if 'checkpoint' in d]
+            for cp in checkpoints:
+                cp['file'] = file
+                cp['file_index'] = i
+                all_checkpoints.append(cp)
+                
+            # Show progress for this file
+            if timeline:
+                start_rss = timeline[0]['rss_gb']
+                end_rss = timeline[-1]['rss_gb']
+                growth = end_rss - start_rss
+                samples = len(timeline)
+                
+                print(f"📊 File {i+1:2d}: {start_rss:5.1f}GB → {end_rss:5.1f}GB "
+                      f"(+{growth:4.1f}GB) [{samples:3d} samples]")
+                
+                # Show significant checkpoints from this file
+                if checkpoints:
+                    for cp in checkpoints:
+                        print(f"    📍 {cp['checkpoint']}: {cp['rss_gb']:.1f}GB")
+                        
+        except Exception as e:
+            print(f"❌ Error reading {file}: {e}")
+    
+    print("\n" + "=" * 60)
+    print("🎯 ANALYSIS SUMMARY")
+    print("=" * 60)
+    
+    print(f"📈 Peak Memory: {peak_memory:.1f} GB")
+    print(f"🎮 Peak VRAM: {peak_vram:.1f} GB")
+    print(f"⚡ Growth Spikes: {len(critical_points)} events >3GB")
+    
+    if critical_points:
+        print(f"\n💥 MEMORY GROWTH SPIKES (>3GB):")
+        print("   Location                           Growth   Total   VRAM")
+        print("   " + "-" * 55)
+        
+        for point in critical_points:
+            location = point['checkpoint'][:30].ljust(30)
+            print(f"   {location} +{point['growth_gb']:4.1f}GB → {point['rss_gb']:5.1f}GB  {point['vram_gb']:4.1f}GB")
+    
+    if all_checkpoints:
+        print(f"\n📍 CHECKPOINT PROGRESSION:")
+        print("   Checkpoint                         Memory   VRAM    File")
+        print("   " + "-" * 55)
+        
+        for cp in all_checkpoints:
+            checkpoint = cp['checkpoint'][:30].ljust(30)
+            file_num = cp['file_index'] + 1
+            print(f"   {checkpoint} {cp['rss_gb']:5.1f}GB  {cp['vram_gb']:4.1f}GB  #{file_num}")
+    
+    # Memory growth analysis
+    if len(all_checkpoints) > 1:
+        print(f"\n📊 MEMORY GROWTH ANALYSIS:")
+        
+        # Find the biggest memory jumps between checkpoints
+        big_jumps = []
+        for i in range(1, len(all_checkpoints)):
+            prev_cp = all_checkpoints[i-1]
+            curr_cp = all_checkpoints[i]
+            
+            growth = curr_cp['rss_gb'] - prev_cp['rss_gb']
+            if growth > 2.0:  # >2GB jump
+                big_jumps.append({
+                    'from': prev_cp['checkpoint'],
+                    'to': curr_cp['checkpoint'],
+                    'growth': growth,
+                    'from_memory': prev_cp['rss_gb'],
+                    'to_memory': curr_cp['rss_gb']
+                })
+        
+        if big_jumps:
+            print("   Major jumps (>2GB):")
+            for jump in big_jumps:
+                print(f"   {jump['from']} → {jump['to']}: "
+                      f"+{jump['growth']:.1f}GB ({jump['from_memory']:.1f}→{jump['to_memory']:.1f}GB)")
+        else:
+            print("   ✅ No major memory jumps detected")
+    
+    # Diagnosis
+    print(f"\n🔬 DIAGNOSIS:")
+    
+    if peak_memory > 400:
+        print("   🔴 CRITICAL: Memory usage exceeded 400GB")
+        print("   💡 Recommendation: Reduce chunk_size to 200-300 frames")
+    elif peak_memory > 200:
+        print("   🟡 HIGH: Memory usage over 200GB")
+        print("   💡 Recommendation: Reduce chunk_size to 400 frames")
+    else:
+        print("   🟢 MODERATE: Memory usage under 200GB")
+    
+    if critical_points:
+        # Find most common growth spike locations
+        spike_locations = {}
+        for point in critical_points:
+            location = point['checkpoint']
+            spike_locations[location] = spike_locations.get(location, 0) + 1
+        
+        print("\n   🎯 Most problematic locations:")
+        for location, count in sorted(spike_locations.items(), key=lambda x: x[1], reverse=True)[:3]:
+            print(f"      {location}: {count} spikes")
+    
+    print(f"\n💡 NEXT STEPS:")
+    if 'merge' in str(critical_points).lower():
+        print("   1. Chunk merging still causing memory accumulation")
+        print("   2. Check if streaming merge is actually being used")
+        print("   3. Verify chunk files are being deleted immediately")
+    elif 'propagation' in str(critical_points).lower():
+        print("   1. SAM2 propagation using too much memory")
+        print("   2. Reduce chunk_size further (try 300 frames)")
+        print("   3. Enable more aggressive frame release")
+    else:
+        print("   1. Review the checkpoint progression above")
+        print("   2. Focus on locations with biggest memory spikes")
+        print("   3. Consider reducing chunk_size if spikes are large")
+
+def main():
+    print("🔍 MEMORY PROFILE ANALYZER")
+    print("Analyzing memory profile files for OOM causes...")
+    print()
+    
+    analyze_memory_files()
+
+if __name__ == "__main__":
+    main()

config_runpod.yaml

@@ -3,8 +3,8 @@ input:
   
 processing:
   scale_factor: 0.5  # A40 can handle 0.5 well
-  chunk_size: 0     # Auto-calculate based on A40's 48GB VRAM
-  overlap_frames: 60
+  chunk_size: 600   # Category A.4: Larger chunks for better VRAM utilization (was 200)
+  overlap_frames: 30 # Reduced overlap
   
 detection:
   confidence_threshold: 0.7
@@ -19,9 +19,11 @@ matting:
   
 output:
   path: "/workspace/output/matted_video.mp4"
-  format: "alpha"
+  format: "greenscreen"  # Changed to greenscreen for easier testing
   background_color: [0, 255, 0]
   maintain_sbs: true
+  preserve_audio: true  # Category A.1: Audio preservation
+  verify_sync: true     # Category A.2: Frame count validation
   
 hardware:
   device: "cuda"

debug_memory_leak.py

@@ -0,0 +1,151 @@
+#!/usr/bin/env python3
+"""
+Debug memory leak between chunks - track exactly where memory accumulates
+"""
+
+import psutil
+import gc
+from pathlib import Path
+import sys
+
+def detailed_memory_check(label):
+    """Get detailed memory info"""
+    process = psutil.Process()
+    memory_info = process.memory_info()
+    
+    rss_gb = memory_info.rss / (1024**3)
+    vms_gb = memory_info.vms / (1024**3)
+    
+    # System memory
+    sys_memory = psutil.virtual_memory()
+    available_gb = sys_memory.available / (1024**3)
+    
+    print(f"🔍 {label}:")
+    print(f"   RSS: {rss_gb:.2f} GB (physical memory)")
+    print(f"   VMS: {vms_gb:.2f} GB (virtual memory)")
+    print(f"   Available: {available_gb:.2f} GB")
+    
+    return rss_gb
+
+def simulate_chunk_processing():
+    """Simulate the chunk processing to see where memory accumulates"""
+    
+    print("🚀 SIMULATING CHUNK PROCESSING TO FIND MEMORY LEAK")
+    print("=" * 60)
+    
+    base_memory = detailed_memory_check("0. Baseline")
+    
+    # Step 1: Import everything (with lazy loading)
+    print("\n📦 Step 1: Imports")
+    from vr180_matting.config import VR180Config
+    from vr180_matting.vr180_processor import VR180Processor
+    
+    import_memory = detailed_memory_check("1. After imports")
+    import_growth = import_memory - base_memory
+    print(f"   Growth: +{import_growth:.2f} GB")
+    
+    # Step 2: Load config
+    print("\n⚙️  Step 2: Config loading")
+    config = VR180Config.from_yaml('config.yaml')
+    config_memory = detailed_memory_check("2. After config load")
+    config_growth = config_memory - import_memory
+    print(f"   Growth: +{config_growth:.2f} GB")
+    
+    # Step 3: Initialize processor (models still lazy)
+    print("\n🏗️  Step 3: Processor initialization")
+    processor = VR180Processor(config)
+    processor_memory = detailed_memory_check("3. After processor init")
+    processor_growth = processor_memory - config_memory
+    print(f"   Growth: +{processor_growth:.2f} GB")
+    
+    # Step 4: Load video info (lightweight)
+    print("\n🎬 Step 4: Video info loading")
+    try:
+        video_info = processor.load_video_info(config.input.video_path)
+        print(f"   Video: {video_info.get('width', 'unknown')}x{video_info.get('height', 'unknown')}, "
+              f"{video_info.get('total_frames', 'unknown')} frames")
+    except Exception as e:
+        print(f"   Warning: Could not load video info: {e}")
+    
+    video_info_memory = detailed_memory_check("4. After video info")
+    video_info_growth = video_info_memory - processor_memory
+    print(f"   Growth: +{video_info_growth:.2f} GB")
+    
+    # Step 5: Simulate chunk 0 processing (this is where models actually load)
+    print("\n🔄 Step 5: Simulating chunk 0 processing...")
+    
+    # This is where the real memory usage starts
+    print("   Loading first 10 frames to trigger model loading...")
+    try:
+        # Read a small number of frames to trigger model loading
+        frames = processor.read_video_frames(
+            config.input.video_path,
+            start_frame=0,
+            num_frames=10,  # Just 10 frames to trigger model loading
+            scale_factor=config.processing.scale_factor
+        )
+        
+        frames_memory = detailed_memory_check("5a. After reading 10 frames")
+        frames_growth = frames_memory - video_info_memory
+        print(f"   10 frames growth: +{frames_growth:.2f} GB")
+        
+        # Free frames
+        del frames
+        gc.collect()
+        
+        after_free_memory = detailed_memory_check("5b. After freeing 10 frames")
+        free_improvement = frames_memory - after_free_memory
+        print(f"   Memory freed: -{free_improvement:.2f} GB")
+        
+    except Exception as e:
+        print(f"   Could not simulate frame loading: {e}")
+        after_free_memory = video_info_memory
+    
+    print(f"\n📊 MEMORY ANALYSIS:")
+    print(f"   Baseline → Final: {base_memory:.2f}GB → {after_free_memory:.2f}GB")
+    print(f"   Total growth: +{after_free_memory - base_memory:.2f}GB")
+    
+    if after_free_memory - base_memory > 10:
+        print(f"   🔴 HIGH: Memory growth > 10GB before any real processing")
+        print(f"   💡 This suggests model loading is using too much memory")
+    elif after_free_memory - base_memory > 5:
+        print(f"   🟡 MODERATE: Memory growth 5-10GB")
+        print(f"   💡 Normal for model loading, but monitor chunk processing")
+    else:
+        print(f"   🟢 GOOD: Memory growth < 5GB")
+        print(f"   💡 Initialization memory usage is reasonable")
+    
+    print(f"\n🎯 KEY INSIGHTS:")
+    if import_growth > 1:
+        print(f"   - Import growth: {import_growth:.2f}GB (fixed with lazy loading)")
+    if processor_growth > 10:
+        print(f"   - Processor init: {processor_growth:.2f}GB (investigate model pre-loading)")
+    
+    print(f"\n💡 RECOMMENDATIONS:")
+    if after_free_memory - base_memory > 15:
+        print(f"   1. Reduce chunk_size to 200-300 frames")
+        print(f"   2. Use smaller models (yolov8n instead of yolov8m)")
+        print(f"   3. Enable FP16 mode for SAM2")
+    elif after_free_memory - base_memory > 8:
+        print(f"   1. Monitor chunk processing carefully")
+        print(f"   2. Use streaming merge (should be automatic)")
+        print(f"   3. Current settings may be acceptable")
+    else:
+        print(f"   1. Settings look good for initialization")
+        print(f"   2. Focus on chunk processing memory leaks")
+
+def main():
+    if len(sys.argv) != 2:
+        print("Usage: python debug_memory_leak.py <config.yaml>")
+        print("This simulates initialization to find memory leaks")
+        sys.exit(1)
+    
+    config_path = sys.argv[1]
+    if not Path(config_path).exists():
+        print(f"Config file not found: {config_path}")
+        sys.exit(1)
+    
+    simulate_chunk_processing()
+
+if __name__ == "__main__":
+    main()

memory_profiler_script.py

@@ -0,0 +1,249 @@
+#!/usr/bin/env python3
+"""
+Memory profiling script for VR180 matting pipeline
+Tracks memory usage during processing to identify leaks
+"""
+
+import sys
+import time
+import psutil
+import tracemalloc
+import subprocess
+import gc
+from pathlib import Path
+from typing import Dict, List, Tuple
+import threading
+import json
+
+class MemoryProfiler:
+    def __init__(self, output_file: str = "memory_profile.json"):
+        self.output_file = output_file
+        self.data = []
+        self.process = psutil.Process()
+        self.running = False
+        self.thread = None
+        self.checkpoint_counter = 0
+        
+    def start_monitoring(self, interval: float = 1.0):
+        """Start continuous memory monitoring"""
+        tracemalloc.start()
+        self.running = True
+        self.thread = threading.Thread(target=self._monitor_loop, args=(interval,))
+        self.thread.daemon = True
+        self.thread.start()
+        print(f"🔍 Memory monitoring started (interval: {interval}s)")
+        
+    def stop_monitoring(self):
+        """Stop monitoring and save results"""
+        self.running = False
+        if self.thread:
+            self.thread.join()
+        
+        # Get tracemalloc snapshot
+        snapshot = tracemalloc.take_snapshot()
+        top_stats = snapshot.statistics('lineno')
+        
+        # Save detailed results
+        results = {
+            'timeline': self.data,
+            'top_memory_allocations': [
+                {
+                    'file': stat.traceback.format()[0],
+                    'size_mb': stat.size / 1024 / 1024,
+                    'count': stat.count
+                }
+                for stat in top_stats[:20]  # Top 20 allocations
+            ],
+            'summary': {
+                'peak_rss_gb': max([d['rss_gb'] for d in self.data]) if self.data else 0,
+                'peak_vram_gb': max([d['vram_gb'] for d in self.data]) if self.data else 0,
+                'total_samples': len(self.data)
+            }
+        }
+        
+        with open(self.output_file, 'w') as f:
+            json.dump(results, f, indent=2)
+        
+        tracemalloc.stop()
+        print(f"📊 Memory profile saved to {self.output_file}")
+        
+    def _monitor_loop(self, interval: float):
+        """Continuous monitoring loop"""
+        while self.running:
+            try:
+                # System memory
+                memory_info = self.process.memory_info()
+                rss_gb = memory_info.rss / (1024**3)
+                
+                # System-wide memory
+                sys_memory = psutil.virtual_memory()
+                sys_used_gb = (sys_memory.total - sys_memory.available) / (1024**3)
+                sys_available_gb = sys_memory.available / (1024**3)
+                
+                # GPU memory (if available)
+                vram_gb = 0
+                vram_free_gb = 0
+                try:
+                    result = subprocess.run(['nvidia-smi', '--query-gpu=memory.used,memory.free', 
+                                           '--format=csv,noheader,nounits'], 
+                                          capture_output=True, text=True, timeout=5)
+                    if result.returncode == 0:
+                        lines = result.stdout.strip().split('\n')
+                        if lines and lines[0]:
+                            used, free = lines[0].split(', ')
+                            vram_gb = float(used) / 1024
+                            vram_free_gb = float(free) / 1024
+                except Exception:
+                    pass
+                
+                # Tracemalloc current usage
+                try:
+                    current, peak = tracemalloc.get_traced_memory()
+                    traced_mb = current / (1024**2)
+                except Exception:
+                    traced_mb = 0
+                
+                data_point = {
+                    'timestamp': time.time(),
+                    'rss_gb': rss_gb,
+                    'vram_gb': vram_gb,
+                    'vram_free_gb': vram_free_gb,
+                    'sys_used_gb': sys_used_gb,
+                    'sys_available_gb': sys_available_gb,
+                    'traced_mb': traced_mb
+                }
+                
+                self.data.append(data_point)
+                
+                # Print periodic updates and save partial data
+                if len(self.data) % 10 == 0:  # Every 10 samples
+                    print(f"🔍 Memory: RSS={rss_gb:.2f}GB, VRAM={vram_gb:.2f}GB, Sys={sys_used_gb:.1f}GB")
+                    
+                    # Save partial data every 30 samples in case of crash
+                    if len(self.data) % 30 == 0:
+                        self._save_partial_data()
+                    
+            except Exception as e:
+                print(f"Monitoring error: {e}")
+                
+            time.sleep(interval)
+    
+    def _save_partial_data(self):
+        """Save partial data to prevent loss on crash"""
+        try:
+            partial_file = f"memory_profile_partial_{self.checkpoint_counter}.json"
+            with open(partial_file, 'w') as f:
+                json.dump({
+                    'timeline': self.data,
+                    'status': 'partial_save',
+                    'samples': len(self.data)
+                }, f, indent=2)
+            self.checkpoint_counter += 1
+        except Exception as e:
+            print(f"Failed to save partial data: {e}")
+    
+    def log_checkpoint(self, checkpoint_name: str):
+        """Log a specific checkpoint"""
+        if self.data:
+            self.data[-1]['checkpoint'] = checkpoint_name
+            latest = self.data[-1]
+            print(f"📍 CHECKPOINT [{checkpoint_name}]: RSS={latest['rss_gb']:.2f}GB, VRAM={latest['vram_gb']:.2f}GB")
+            
+            # Save checkpoint data immediately
+            self._save_partial_data()
+
+def run_with_profiling(config_path: str):
+    """Run the VR180 matting with memory profiling"""
+    profiler = MemoryProfiler("memory_profile_detailed.json")
+    
+    try:
+        # Start monitoring
+        profiler.start_monitoring(interval=2.0)  # Sample every 2 seconds
+        
+        # Log initial state
+        profiler.log_checkpoint("STARTUP")
+        
+        # Import after starting profiler to catch import memory usage
+        print("Importing VR180 processor...")
+        from vr180_matting.vr180_processor import VR180Processor
+        from vr180_matting.config import VR180Config
+        
+        profiler.log_checkpoint("IMPORTS_COMPLETE")
+        
+        # Load config
+        print(f"Loading config from {config_path}")
+        config = VR180Config.from_yaml(config_path)
+        
+        profiler.log_checkpoint("CONFIG_LOADED")
+        
+        # Initialize processor
+        print("Initializing VR180 processor...")
+        processor = VR180Processor(config)
+        
+        profiler.log_checkpoint("PROCESSOR_INITIALIZED")
+        
+        # Force garbage collection
+        gc.collect()
+        profiler.log_checkpoint("INITIAL_GC_COMPLETE")
+        
+        # Run processing
+        print("Starting VR180 processing...")
+        processor.process_video()
+        
+        profiler.log_checkpoint("PROCESSING_COMPLETE")
+        
+    except Exception as e:
+        print(f"❌ Error during processing: {e}")
+        profiler.log_checkpoint(f"ERROR: {str(e)}")
+        raise
+    finally:
+        # Stop monitoring and save results
+        profiler.stop_monitoring()
+        
+        # Print summary
+        print("\n" + "="*60)
+        print("MEMORY PROFILING SUMMARY")
+        print("="*60)
+        
+        if profiler.data:
+            peak_rss = max([d['rss_gb'] for d in profiler.data])
+            peak_vram = max([d['vram_gb'] for d in profiler.data])
+            
+            print(f"Peak RSS Memory: {peak_rss:.2f} GB")
+            print(f"Peak VRAM Usage: {peak_vram:.2f} GB")
+            print(f"Total Samples: {len(profiler.data)}")
+            
+            # Show checkpoints
+            checkpoints = [d for d in profiler.data if 'checkpoint' in d]
+            if checkpoints:
+                print(f"\nCheckpoints ({len(checkpoints)}):")
+                for cp in checkpoints:
+                    print(f"  {cp['checkpoint']}: RSS={cp['rss_gb']:.2f}GB, VRAM={cp['vram_gb']:.2f}GB")
+        
+        print(f"\nDetailed profile saved to: {profiler.output_file}")
+
+def main():
+    if len(sys.argv) != 2:
+        print("Usage: python memory_profiler_script.py <config.yaml>")
+        print("\nThis script runs VR180 matting with detailed memory profiling")
+        print("It will:")
+        print("- Monitor RSS, VRAM, and system memory every 2 seconds")
+        print("- Track memory allocations with tracemalloc")
+        print("- Log checkpoints at key processing stages")
+        print("- Save detailed JSON report for analysis")
+        sys.exit(1)
+    
+    config_path = sys.argv[1]
+    
+    if not Path(config_path).exists():
+        print(f"❌ Config file not found: {config_path}")
+        sys.exit(1)
+    
+    print("🚀 Starting VR180 Memory Profiling")
+    print(f"Config: {config_path}")
+    print("="*60)
+    
+    run_with_profiling(config_path)
+
+if __name__ == "__main__":
+    main()

quick_memory_check.py

@@ -0,0 +1,125 @@
+#!/usr/bin/env python3
+"""
+Quick memory and system check before running full pipeline
+"""
+
+import psutil
+import subprocess
+import sys
+from pathlib import Path
+
+def check_system():
+    """Check system resources before starting"""
+    print("🔍 SYSTEM RESOURCE CHECK")
+    print("=" * 50)
+    
+    # Memory info
+    memory = psutil.virtual_memory()
+    print(f"📊 RAM:")
+    print(f"   Total: {memory.total / (1024**3):.1f} GB")
+    print(f"   Available: {memory.available / (1024**3):.1f} GB")
+    print(f"   Used: {(memory.total - memory.available) / (1024**3):.1f} GB ({memory.percent:.1f}%)")
+    
+    # GPU info
+    try:
+        result = subprocess.run(['nvidia-smi', '--query-gpu=name,memory.total,memory.used,memory.free', 
+                               '--format=csv,noheader,nounits'], 
+                              capture_output=True, text=True, timeout=10)
+        if result.returncode == 0:
+            lines = result.stdout.strip().split('\n')
+            print(f"\n🎮 GPU:")
+            for i, line in enumerate(lines):
+                if line.strip():
+                    parts = line.split(', ')
+                    if len(parts) >= 4:
+                        name, total, used, free = parts[:4]
+                        total_gb = float(total) / 1024
+                        used_gb = float(used) / 1024
+                        free_gb = float(free) / 1024
+                        print(f"   GPU {i}: {name}")
+                        print(f"   VRAM: {used_gb:.1f}/{total_gb:.1f} GB ({used_gb/total_gb*100:.1f}% used)")
+                        print(f"   Free: {free_gb:.1f} GB")
+    except Exception as e:
+        print(f"\n⚠️  Could not get GPU info: {e}")
+    
+    # Disk space
+    disk = psutil.disk_usage('/')
+    print(f"\n💾 Disk (/):")
+    print(f"   Total: {disk.total / (1024**3):.1f} GB")
+    print(f"   Used: {disk.used / (1024**3):.1f} GB ({disk.used/disk.total*100:.1f}%)")
+    print(f"   Free: {disk.free / (1024**3):.1f} GB")
+    
+    # Check config file
+    if len(sys.argv) > 1:
+        config_path = sys.argv[1]
+        if Path(config_path).exists():
+            print(f"\n✅ Config file found: {config_path}")
+            
+            # Try to load and show key settings
+            try:
+                import yaml
+                with open(config_path, 'r') as f:
+                    config = yaml.safe_load(f)
+                
+                print(f"📋 Key Settings:")
+                if 'processing' in config:
+                    proc = config['processing']
+                    print(f"   Chunk size: {proc.get('chunk_size', 'default')}")
+                    print(f"   Scale factor: {proc.get('scale_factor', 'default')}")
+                
+                if 'hardware' in config:
+                    hw = config['hardware']
+                    print(f"   Max VRAM: {hw.get('max_vram_gb', 'default')} GB")
+                    
+                if 'input' in config:
+                    inp = config['input']
+                    video_path = inp.get('video_path', '')
+                    if video_path and Path(video_path).exists():
+                        size_gb = Path(video_path).stat().st_size / (1024**3)
+                        print(f"   Input video: {video_path} ({size_gb:.1f} GB)")
+                    else:
+                        print(f"   ⚠️  Input video not found: {video_path}")
+                        
+            except Exception as e:
+                print(f"   ⚠️  Could not parse config: {e}")
+        else:
+            print(f"\n❌ Config file not found: {config_path}")
+            return False
+    
+    # Memory safety warnings
+    print(f"\n⚠️  MEMORY SAFETY CHECKS:")
+    available_gb = memory.available / (1024**3)
+    
+    if available_gb < 10:
+        print(f"   🔴 LOW MEMORY: Only {available_gb:.1f}GB available")
+        print("      Consider: reducing chunk_size or scale_factor")
+        return False
+    elif available_gb < 20:
+        print(f"   🟡 MODERATE MEMORY: {available_gb:.1f}GB available")
+        print("      Recommend: chunk_size ≤ 300, scale_factor ≤ 0.5")
+    else:
+        print(f"   🟢 GOOD MEMORY: {available_gb:.1f}GB available")
+    
+    print(f"\n" + "=" * 50)
+    return True
+
+def main():
+    if len(sys.argv) != 2:
+        print("Usage: python quick_memory_check.py <config.yaml>")
+        print("\nThis checks system resources before running VR180 matting")
+        sys.exit(1)
+    
+    safe_to_run = check_system()
+    
+    if safe_to_run:
+        print("✅ System check passed - safe to run VR180 matting")
+        print("\nTo run with memory profiling:")
+        print(f"   python memory_profiler_script.py {sys.argv[1]}")
+        print("\nTo run normally:")
+        print(f"   vr180-matting {sys.argv[1]}")
+    else:
+        print("❌ System check failed - address issues before running")
+        sys.exit(1)
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -9,3 +9,7 @@ ultralytics>=8.0.0
 tqdm>=4.65.0
 psutil>=5.9.0
 ffmpeg-python>=0.2.0
+decord>=0.6.0
+# GPU acceleration (optional but recommended for stereo validation speedup)
+# cupy-cuda11x>=12.0.0  # For CUDA 11.x
+# cupy-cuda12x>=12.0.0  # For CUDA 12.x - uncomment appropriate version

runpod_setup.sh

@@ -14,6 +14,32 @@ echo "🐍 Installing Python dependencies..."
 pip install --upgrade pip
 pip install -r requirements.txt
 
+# Install decord for SAM2 video loading
+echo "📹 Installing decord for video processing..."
+pip install decord
+
+# Install CuPy for GPU acceleration of stereo validation
+echo "🚀 Installing CuPy for GPU acceleration..."
+# Auto-detect CUDA version and install appropriate CuPy
+python -c "
+import torch
+if torch.cuda.is_available():
+    cuda_version = torch.version.cuda
+    print(f'CUDA version detected: {cuda_version}')
+    if cuda_version.startswith('11.'):
+        import subprocess
+        subprocess.run(['pip', 'install', 'cupy-cuda11x>=12.0.0'])
+        print('Installed CuPy for CUDA 11.x')
+    elif cuda_version.startswith('12.'):
+        import subprocess
+        subprocess.run(['pip', 'install', 'cupy-cuda12x>=12.0.0'])
+        print('Installed CuPy for CUDA 12.x')
+    else:
+        print(f'Unsupported CUDA version: {cuda_version}')
+else:
+    print('CUDA not available, skipping CuPy installation')
+"
+
 # Install SAM2 separately (not on PyPI)
 echo "🎯 Installing SAM2..."
 pip install git+https://github.com/facebookresearch/segment-anything-2.git

spec.md

@@ -123,6 +123,204 @@ hardware:
 3. **Performance Profiling**: Detailed resource usage analytics
 4. **Quality Validation**: Comprehensive testing suite
 
+## Post-Implementation Optimization Opportunities
+
+*Based on first successful 30-second test clip execution results (A40 GPU, 50% scale, 9x200 frame chunks)*
+
+### Performance Analysis Findings
+- **Processing Speed**: ~0.54s per frame (64.4s for 120 frames per chunk)
+- **VRAM Utilization**: Only 2.5% (1.11GB of 45GB available) - significantly underutilized
+- **RAM Usage**: 106GB used of 494GB available (21.5%)
+- **Primary Bottleneck**: Intermediate ffmpeg encoding operations per chunk
+
+### Identified Optimization Categories
+
+#### Category A: Performance Improvements (Quick Wins)
+1. **Audio Track Preservation** ⚠️ **CRITICAL**
+   - Issue: Output video missing audio track from input
+   - Solution: Use ffmpeg to copy audio stream during final video creation
+   - Implementation: Add `-c:a copy` to final ffmpeg command
+   - Impact: Essential for production usability
+   - Risk: Low, standard ffmpeg operation
+
+2. **Frame Count Synchronization** ⚠️ **CRITICAL**
+   - Issue: Audio sync drift if input/output frame counts differ
+   - Solution: Validate exact frame count preservation throughout pipeline
+   - Implementation: Frame count verification + duration matching
+   - Impact: Prevents audio desync in long videos
+   - Risk: Low, validation feature
+
+3. **Memory Usage Reality Check** ⚠️ **IMPORTANT**
+   - Current assumption: Unlimited RAM for memory-only pipeline
+   - Reality: RunPod container limited to ~48GB RAM
+   - Risk calculation: 1-hour video = ~213k frames = potential 20-40GB+ memory usage
+   - Solution: Implement streaming output instead of full in-memory accumulation
+   - Impact: Enables processing of long-form content
+   - Risk: Medium, requires pipeline restructuring
+
+4. **Larger Chunk Sizes**
+   - Current: 200 frames per chunk (conservative for 10GB RTX 3080)
+   - Opportunity: 600-800 frames per chunk on high-VRAM systems
+   - Impact: Reduce 9 chunks to 2-3 chunks, fewer intermediate operations
+   - Risk: Low, easily configurable
+
+5. **Streaming Output Pipeline**
+   - Current: Accumulate all processed frames in memory, write once
+   - Opportunity: Write processed chunks to temporary segments, merge at end
+   - Impact: Constant memory usage regardless of video length
+   - Risk: Medium, requires temporary file management
+
+6. **Enhanced Performance Profiling**
+   - Current: Basic memory monitoring
+   - Opportunity: Detailed timing per processing stage (detection, propagation, encoding)
+   - Impact: Identify exact bottlenecks for targeted optimization
+   - Risk: Low, debugging feature
+
+7. **Parallel Eye Processing**
+   - Current: Sequential left eye → right eye processing
+   - Opportunity: Process both eyes simultaneously
+   - Impact: Potential 50% speedup, better GPU utilization
+   - Risk: Medium, memory management complexity
+
+#### Category B: Stereo Consistency Fixes (Critical for VR)
+1. **Master-Slave Eye Processing**
+   - Issue: Independent detection leads to mismatched person counts between eyes
+   - Solution: Use left eye detections as "seeds" for right eye processing
+   - Impact: Ensures identical person detection across stereo pair
+   - Risk: Low, maintains current quality while improving consistency
+
+2. **Cross-Eye Detection Validation**
+   - Issue: Hair/clothing included on one eye but not the other
+   - Solution: Compare detection results, flag inconsistencies for reprocessing
+   - Impact: 90%+ stereo alignment improvement
+   - Risk: Low, fallback to current behavior
+
+3. **Disparity-Aware Segmentation**
+   - Issue: Segmentation boundaries differ between eyes despite same person
+   - Solution: Use stereo disparity to correlate features between eyes
+   - Impact: True stereo-consistent matting
+   - Risk: High, complex implementation
+
+4. **Joint Stereo Detection**
+   - Issue: YOLO runs independently on each eye
+   - Solution: Run YOLO on full SBS frame, split detections spatially
+   - Impact: Guaranteed identical detection counts
+   - Risk: Medium, requires detection coordinate mapping
+
+#### Category C: Advanced Optimizations (Future)
+1. **Adaptive Memory Management**
+   - Opportunity: Dynamic chunk sizing based on real-time VRAM usage
+   - Impact: Optimal resource utilization across different hardware
+   - Risk: Medium, complex heuristics
+
+2. **Multi-Resolution Processing**
+   - Opportunity: Initial processing at lower resolution, edge refinement at full
+   - Impact: Speed improvement while maintaining quality
+   - Risk: Medium, quality validation required
+
+3. **Enhanced Workflow Documentation**
+   - Issue: Unclear intermediate data lifecycle
+   - Solution: Detailed logging of chunk processing, optional intermediate preservation
+   - Impact: Better debugging and user understanding
+   - Risk: Low, documentation feature
+
+### Implementation Strategy
+- **Phase A**: Quick performance wins (larger chunks, profiling)
+- **Phase B**: Stereo consistency (master-slave, validation)  
+- **Phase C**: Advanced features (disparity-aware, memory optimization)
+
+### Configuration Extensions Required
+```yaml
+processing:
+  chunk_size: 600  # Increase from 200 for high-VRAM systems
+  memory_pipeline: false  # Skip intermediate video creation (disabled due to RAM limits)
+  streaming_output: true  # Write chunks progressively instead of accumulating
+  parallel_eyes: false  # Process eyes simultaneously
+  max_memory_gb: 40  # Realistic RAM limit for RunPod containers
+  
+audio:
+  preserve_audio: true  # Copy audio track from input to output
+  verify_sync: true  # Validate frame count and duration matching
+  audio_codec: "copy"  # Preserve original audio codec
+  
+stereo:
+  consistency_mode: "master_slave"  # "independent", "master_slave", "joint"
+  validation_threshold: 0.8  # Similarity threshold between eyes
+  correction_method: "transfer"  # "transfer", "reprocess", "ensemble"
+  
+performance:
+  profile_enabled: true  # Detailed timing analysis
+  preserve_intermediates: false  # For debugging workflow
+
+debugging:
+  log_intermediate_workflow: true  # Document chunk lifecycle
+  save_detection_visualization: false  # Debug detection mismatches
+  frame_count_validation: true  # Ensure exact frame preservation
+```
+
+### Technical Implementation Details
+
+#### Audio Preservation Implementation
+```python
+# During final video save, include audio stream copy
+ffmpeg_cmd = [
+    'ffmpeg', '-y',
+    '-framerate', str(fps),
+    '-i', frame_pattern,           # Video frames
+    '-i', input_video_path,        # Original video for audio
+    '-c:v', 'h264_nvenc',         # GPU video codec (with CPU fallback)
+    '-c:a', 'copy',               # Copy audio without re-encoding
+    '-map', '0:v:0',              # Map video from first input
+    '-map', '1:a:0',              # Map audio from second input
+    '-shortest',                  # Match shortest stream duration
+    output_path
+]
+```
+
+#### Streaming Output Implementation
+```python
+# Instead of accumulating frames in memory:
+class StreamingVideoWriter:
+    def __init__(self, output_path, fps, audio_source):
+        self.temp_segments = []
+        self.current_segment = 0
+        
+    def write_chunk(self, processed_frames):
+        # Write chunk to temporary segment
+        segment_path = f"temp_segment_{self.current_segment}.mp4"
+        self.write_video_segment(processed_frames, segment_path)
+        self.temp_segments.append(segment_path)
+        self.current_segment += 1
+        
+    def finalize(self):
+        # Merge all segments with audio preservation
+        self.merge_segments_with_audio()
+```
+
+#### Memory Usage Calculation
+```python
+def estimate_memory_requirements(duration_seconds, fps, resolution_scale=0.5):
+    """Calculate memory usage for different video lengths"""
+    frames = duration_seconds * fps
+    
+    # Per-frame memory (rough estimates for VR180 at 50% scale)
+    frame_size_mb = (3072 * 1536 * 3 * 4) / (1024 * 1024)  # ~18MB per frame
+    
+    total_memory_gb = (frames * frame_size_mb) / 1024
+    
+    return {
+        'duration': duration_seconds,
+        'total_frames': frames,
+        'estimated_memory_gb': total_memory_gb,
+        'safe_for_48gb': total_memory_gb < 40
+    }
+
+# Example outputs:
+# 30 seconds: ~2.7GB (safe)
+# 5 minutes: ~27GB (borderline) 
+# 1 hour: ~324GB (requires streaming)
+```
+
 ## Success Criteria
 
 ### Technical Feasibility

test_inter_chunk_cleanup.py

@@ -0,0 +1,148 @@
+#!/usr/bin/env python3
+"""
+Test script to verify inter-chunk cleanup properly destroys models
+"""
+
+import psutil
+import gc
+import sys
+from pathlib import Path
+
+def get_memory_usage():
+    """Get current memory usage in GB"""
+    process = psutil.Process()
+    return process.memory_info().rss / (1024**3)
+
+def test_inter_chunk_cleanup():
+    """Test that models are properly destroyed between chunks"""
+    
+    print("🧪 TESTING INTER-CHUNK CLEANUP")
+    print("=" * 50)
+    
+    baseline_memory = get_memory_usage()
+    print(f"📊 Baseline memory: {baseline_memory:.2f} GB")
+    
+    # Import and create processor
+    print("\n1️⃣ Creating processor...")
+    from vr180_matting.config import VR180Config
+    from vr180_matting.vr180_processor import VR180Processor
+    
+    config = VR180Config.from_yaml('config.yaml')
+    processor = VR180Processor(config)
+    
+    init_memory = get_memory_usage()
+    print(f"📊 After processor init: {init_memory:.2f} GB (+{init_memory - baseline_memory:.2f} GB)")
+    
+    # Simulate chunk processing (just trigger model loading)
+    print("\n2️⃣ Simulating chunk 0 processing...")
+    
+    # Test 1: Force YOLO model loading
+    try:
+        detector = processor.detector
+        detector._load_model()  # Force load
+        yolo_memory = get_memory_usage()
+        print(f"📊 After YOLO load: {yolo_memory:.2f} GB (+{yolo_memory - init_memory:.2f} GB)")
+    except Exception as e:
+        print(f"❌ YOLO loading failed: {e}")
+        yolo_memory = init_memory
+    
+    # Test 2: Force SAM2 model loading
+    try:
+        sam2_model = processor.sam2_model
+        sam2_model._load_model(sam2_model.model_cfg, sam2_model.checkpoint_path)
+        sam2_memory = get_memory_usage()
+        print(f"📊 After SAM2 load: {sam2_memory:.2f} GB (+{sam2_memory - yolo_memory:.2f} GB)")
+    except Exception as e:
+        print(f"❌ SAM2 loading failed: {e}")
+        sam2_memory = yolo_memory
+    
+    total_model_memory = sam2_memory - init_memory
+    print(f"📊 Total model memory: {total_model_memory:.2f} GB")
+    
+    # Test 3: Inter-chunk cleanup
+    print("\n3️⃣ Testing inter-chunk cleanup...")
+    processor._complete_inter_chunk_cleanup(chunk_idx=0)
+    
+    cleanup_memory = get_memory_usage()
+    cleanup_improvement = sam2_memory - cleanup_memory
+    print(f"📊 After cleanup: {cleanup_memory:.2f} GB (-{cleanup_improvement:.2f} GB freed)")
+    
+    # Test 4: Verify models reload fresh
+    print("\n4️⃣ Testing fresh model reload...")
+    
+    # Check YOLO state
+    yolo_reloaded = processor.detector.model is None
+    print(f"🔍 YOLO model destroyed: {'✅ YES' if yolo_reloaded else '❌ NO'}")
+    
+    # Check SAM2 state  
+    sam2_reloaded = not processor.sam2_model._model_loaded or processor.sam2_model.predictor is None
+    print(f"🔍 SAM2 model destroyed: {'✅ YES' if sam2_reloaded else '❌ NO'}")
+    
+    # Test 5: Force reload to verify they work
+    print("\n5️⃣ Testing model reload...")
+    try:
+        # Force YOLO reload
+        processor.detector._load_model()
+        yolo_reload_memory = get_memory_usage()
+        
+        # Force SAM2 reload
+        processor.sam2_model._load_model(processor.sam2_model.model_cfg, processor.sam2_model.checkpoint_path)
+        sam2_reload_memory = get_memory_usage()
+        
+        reload_growth = sam2_reload_memory - cleanup_memory
+        print(f"📊 After reload: {sam2_reload_memory:.2f} GB (+{reload_growth:.2f} GB)")
+        
+        if abs(reload_growth - total_model_memory) < 1.0:  # Within 1GB
+            print("✅ Models reloaded with similar memory usage (good)")
+        else:
+            print("⚠️  Model reload memory differs significantly")
+            
+    except Exception as e:
+        print(f"❌ Model reload failed: {e}")
+    
+    # Final summary
+    print(f"\n📊 SUMMARY:")
+    print(f"   Baseline → Peak: {baseline_memory:.2f}GB → {sam2_memory:.2f}GB")
+    print(f"   Peak → Cleanup: {sam2_memory:.2f}GB → {cleanup_memory:.2f}GB")
+    print(f"   Memory freed: {cleanup_improvement:.2f}GB")
+    print(f"   Models destroyed: YOLO={yolo_reloaded}, SAM2={sam2_reloaded}")
+    
+    # Success criteria: Both models destroyed AND can reload
+    models_destroyed = yolo_reloaded and sam2_reloaded
+    can_reload = 'reload_growth' in locals()
+    
+    if models_destroyed and can_reload:
+        print("✅ Inter-chunk cleanup working effectively")
+        print("💡 Models destroyed and can reload fresh (memory will be freed during real processing)")
+        return True
+    elif models_destroyed:
+        print("⚠️  Models destroyed but reload test incomplete")
+        print("💡 This should still prevent accumulation during real processing")
+        return True
+    else:
+        print("❌ Inter-chunk cleanup not freeing enough memory")
+        return False
+
+def main():
+    if len(sys.argv) != 2:
+        print("Usage: python test_inter_chunk_cleanup.py <config.yaml>")
+        sys.exit(1)
+    
+    config_path = sys.argv[1]
+    if not Path(config_path).exists():
+        print(f"Config file not found: {config_path}")
+        sys.exit(1)
+    
+    success = test_inter_chunk_cleanup()
+    
+    if success:
+        print(f"\n🎉 SUCCESS: Inter-chunk cleanup is working!")
+        print(f"💡 This should prevent 15-20GB model accumulation between chunks")
+    else:
+        print(f"\n❌ FAILURE: Inter-chunk cleanup needs improvement")
+        print(f"💡 Check model destruction logic in _complete_inter_chunk_cleanup")
+    
+    return 0 if success else 1
+
+if __name__ == "__main__":
+    sys.exit(main())

vr180_matting/checkpoint_manager.py

@@ -0,0 +1,220 @@
+"""
+Checkpoint manager for resumable video processing
+Saves progress to avoid reprocessing after OOM or crashes
+"""
+
+import json
+import hashlib
+from pathlib import Path
+from typing import Dict, Any, Optional, List
+import os
+import shutil
+from datetime import datetime
+
+
+class CheckpointManager:
+    """Manages processing checkpoints for resumable execution"""
+    
+    def __init__(self, video_path: str, output_path: str, checkpoint_dir: Optional[Path] = None):
+        """
+        Initialize checkpoint manager
+        
+        Args:
+            video_path: Input video path
+            output_path: Output video path  
+            checkpoint_dir: Directory for checkpoint files (default: .vr180_checkpoints in CWD)
+        """
+        self.video_path = Path(video_path)
+        self.output_path = Path(output_path)
+        
+        # Create unique checkpoint ID based on video file
+        self.video_hash = self._compute_video_hash()
+        
+        # Setup checkpoint directory
+        if checkpoint_dir is None:
+            self.checkpoint_dir = Path.cwd() / ".vr180_checkpoints" / self.video_hash
+        else:
+            self.checkpoint_dir = Path(checkpoint_dir) / self.video_hash
+            
+        self.checkpoint_dir.mkdir(parents=True, exist_ok=True)
+        
+        # Checkpoint files
+        self.status_file = self.checkpoint_dir / "processing_status.json"
+        self.chunks_dir = self.checkpoint_dir / "chunks"
+        self.chunks_dir.mkdir(exist_ok=True)
+        
+        # Load existing status or create new
+        self.status = self._load_status()
+        
+    def _compute_video_hash(self) -> str:
+        """Compute hash of video file for unique identification"""
+        # Use file path, size, and modification time for quick hash
+        stat = self.video_path.stat()
+        hash_str = f"{self.video_path}_{stat.st_size}_{stat.st_mtime}"
+        return hashlib.md5(hash_str.encode()).hexdigest()[:12]
+    
+    def _load_status(self) -> Dict[str, Any]:
+        """Load processing status from checkpoint file"""
+        if self.status_file.exists():
+            with open(self.status_file, 'r') as f:
+                status = json.load(f)
+                print(f"📋 Loaded checkpoint: {status['completed_chunks']}/{status['total_chunks']} chunks completed")
+                return status
+        else:
+            # Create new status
+            return {
+                'video_path': str(self.video_path),
+                'output_path': str(self.output_path),
+                'video_hash': self.video_hash,
+                'start_time': datetime.now().isoformat(),
+                'total_chunks': 0,
+                'completed_chunks': 0,
+                'chunk_info': {},
+                'processing_complete': False,
+                'merge_complete': False
+            }
+    
+    def _save_status(self):
+        """Save current status to checkpoint file"""
+        self.status['last_update'] = datetime.now().isoformat()
+        with open(self.status_file, 'w') as f:
+            json.dump(self.status, f, indent=2)
+    
+    def set_total_chunks(self, total_chunks: int):
+        """Set total number of chunks to process"""
+        self.status['total_chunks'] = total_chunks
+        self._save_status()
+    
+    def is_chunk_completed(self, chunk_idx: int) -> bool:
+        """Check if a chunk has already been processed"""
+        chunk_key = f"chunk_{chunk_idx}"
+        return chunk_key in self.status['chunk_info'] and \
+               self.status['chunk_info'][chunk_key].get('completed', False)
+    
+    def get_chunk_file(self, chunk_idx: int) -> Optional[Path]:
+        """Get saved chunk file path if it exists"""
+        chunk_file = self.chunks_dir / f"chunk_{chunk_idx:04d}.npz"
+        if chunk_file.exists() and self.is_chunk_completed(chunk_idx):
+            return chunk_file
+        return None
+    
+    def save_chunk(self, chunk_idx: int, frames: List, source_chunk_path: Optional[Path] = None):
+        """
+        Save processed chunk and mark as completed
+        
+        Args:
+            chunk_idx: Chunk index
+            frames: Processed frames (can be None if using source_chunk_path)
+            source_chunk_path: If provided, copy this file instead of saving frames
+        """
+        chunk_file = self.chunks_dir / f"chunk_{chunk_idx:04d}.npz"
+        
+        try:
+            if source_chunk_path and source_chunk_path.exists():
+                # Copy existing chunk file
+                shutil.copy2(source_chunk_path, chunk_file)
+                print(f"💾 Copied chunk {chunk_idx} to checkpoint: {chunk_file.name}")
+            elif frames is not None:
+                # Save new frames
+                import numpy as np
+                np.savez_compressed(str(chunk_file), frames=frames)
+                print(f"💾 Saved chunk {chunk_idx} to checkpoint: {chunk_file.name}")
+            else:
+                raise ValueError("Either frames or source_chunk_path must be provided")
+            
+            # Update status
+            chunk_key = f"chunk_{chunk_idx}"
+            self.status['chunk_info'][chunk_key] = {
+                'completed': True,
+                'file': chunk_file.name,
+                'timestamp': datetime.now().isoformat()
+            }
+            self.status['completed_chunks'] = len([c for c in self.status['chunk_info'].values() if c['completed']])
+            self._save_status()
+            
+            print(f"✅ Chunk {chunk_idx} checkpoint saved ({self.status['completed_chunks']}/{self.status['total_chunks']})")
+            
+        except Exception as e:
+            print(f"❌ Failed to save chunk {chunk_idx} checkpoint: {e}")
+    
+    def get_completed_chunk_files(self) -> List[Path]:
+        """Get list of all completed chunk files in order"""
+        chunk_files = []
+        missing_chunks = []
+        
+        for i in range(self.status['total_chunks']):
+            chunk_file = self.get_chunk_file(i)
+            if chunk_file:
+                chunk_files.append(chunk_file)
+            else:
+                # Check if chunk is marked as completed but file is missing
+                if self.is_chunk_completed(i):
+                    missing_chunks.append(i)
+                    print(f"⚠️  Chunk {i} marked complete but file missing!")
+                else:
+                    break  # Stop at first unprocessed chunk
+        
+        if missing_chunks:
+            print(f"❌ Missing checkpoint files for chunks: {missing_chunks}")
+            print(f"   This may happen if files were deleted during streaming merge")
+            print(f"   These chunks may need to be reprocessed")
+        
+        return chunk_files
+    
+    def mark_processing_complete(self):
+        """Mark all chunk processing as complete"""
+        self.status['processing_complete'] = True
+        self._save_status()
+        print(f"✅ All chunks processed and checkpointed")
+    
+    def mark_merge_complete(self):
+        """Mark final merge as complete"""
+        self.status['merge_complete'] = True
+        self._save_status()
+        print(f"✅ Video merge completed")
+    
+    def cleanup_checkpoints(self, keep_chunks: bool = False):
+        """
+        Clean up checkpoint files after successful completion
+        
+        Args:
+            keep_chunks: If True, keep chunk files but remove status
+        """
+        if keep_chunks:
+            # Just remove status file
+            if self.status_file.exists():
+                self.status_file.unlink()
+                print(f"🗑️  Removed checkpoint status file")
+        else:
+            # Remove entire checkpoint directory
+            if self.checkpoint_dir.exists():
+                shutil.rmtree(self.checkpoint_dir)
+                print(f"🗑️  Removed all checkpoint files: {self.checkpoint_dir}")
+    
+    def get_resume_info(self) -> Dict[str, Any]:
+        """Get information about what can be resumed"""
+        return {
+            'can_resume': self.status['completed_chunks'] > 0,
+            'completed_chunks': self.status['completed_chunks'],
+            'total_chunks': self.status['total_chunks'],
+            'processing_complete': self.status['processing_complete'],
+            'merge_complete': self.status['merge_complete'],
+            'checkpoint_dir': str(self.checkpoint_dir)
+        }
+    
+    def print_status(self):
+        """Print current checkpoint status"""
+        print(f"\n📊 CHECKPOINT STATUS:")
+        print(f"   Video: {self.video_path.name}")
+        print(f"   Hash: {self.video_hash}")
+        print(f"   Progress: {self.status['completed_chunks']}/{self.status['total_chunks']} chunks")
+        print(f"   Processing complete: {self.status['processing_complete']}")
+        print(f"   Merge complete: {self.status['merge_complete']}")
+        print(f"   Checkpoint dir: {self.checkpoint_dir}")
+        
+        if self.status['completed_chunks'] > 0:
+            print(f"\n   Completed chunks:")
+            for i in range(self.status['completed_chunks']):
+                chunk_info = self.status['chunk_info'].get(f'chunk_{i}', {})
+                if chunk_info.get('completed'):
+                    print(f"     ✓ Chunk {i}: {chunk_info.get('file', 'unknown')}")

vr180_matting/config.py

@@ -29,6 +29,11 @@ class MattingConfig:
     fp16: bool = True
     sam2_model_cfg: str = "sam2.1_hiera_l"
     sam2_checkpoint: str = "segment-anything-2/checkpoints/sam2.1_hiera_large.pt"
+    # Det-SAM2 optimizations
+    continuous_correction: bool = True
+    correction_interval: int = 60  # Add correction prompts every N frames
+    frame_release_interval: int = 50  # Release old frames every N frames
+    frame_window_size: int = 30  # Keep N frames in memory
 
 
 @dataclass
@@ -37,6 +42,8 @@ class OutputConfig:
     format: str = "alpha"
     background_color: List[int] = None
     maintain_sbs: bool = True
+    preserve_audio: bool = True
+    verify_sync: bool = True
 
     def __post_init__(self):
         if self.background_color is None:
@@ -99,7 +106,9 @@ class VR180Config:
                 'path': self.output.path,
                 'format': self.output.format,
                 'background_color': self.output.background_color,
-                'maintain_sbs': self.output.maintain_sbs
+                'maintain_sbs': self.output.maintain_sbs,
+                'preserve_audio': self.output.preserve_audio,
+                'verify_sync': self.output.verify_sync
             },
             'hardware': {
                 'device': self.hardware.device,

vr180_matting/detector.py

@@ -1,6 +1,4 @@
-import torch
 import numpy as np
-from ultralytics import YOLO
 from typing import List, Tuple, Dict, Any
 import cv2
 
@@ -13,14 +11,23 @@ class YOLODetector:
         self.confidence_threshold = confidence_threshold
         self.device = device
         self.model = None
-        self._load_model()
+        # Don't load model during init - load lazily when first used
     
     def _load_model(self):
-        """Load YOLOv8 model"""
+        """Load YOLOv8 model lazily"""
+        if self.model is not None:
+            return  # Already loaded
+            
         try:
+            # Import heavy dependencies only when needed
+            import torch
+            from ultralytics import YOLO
+            
             self.model = YOLO(f"{self.model_name}.pt")
             if self.device == "cuda" and torch.cuda.is_available():
                 self.model.to("cuda")
+                
+            print(f"🎯 Loaded YOLO model: {self.model_name}")
         except Exception as e:
             raise RuntimeError(f"Failed to load YOLO model {self.model_name}: {e}")
     
@@ -34,8 +41,9 @@ class YOLODetector:
         Returns:
             List of detection dictionaries with bbox, confidence, and class info
         """
+        # Load model lazily on first use
         if self.model is None:
-            raise RuntimeError("YOLO model not loaded")
+            self._load_model()
         
         results = self.model(frame, verbose=False)
         detections = []

vr180_matting/sam2_wrapper.py

@@ -5,13 +5,20 @@ import cv2
 from pathlib import Path
 import warnings
 import os
+import tempfile
+import shutil
+import gc
 
+# Check SAM2 availability without importing heavy modules
+def _check_sam2_available():
     try:
-    from sam2.build_sam import build_sam2_video_predictor
-    from sam2.sam2_image_predictor import SAM2ImagePredictor
-    SAM2_AVAILABLE = True
+        import sam2
+        return True
     except ImportError:
-    SAM2_AVAILABLE = False
+        return False
+
+SAM2_AVAILABLE = _check_sam2_available()
+if not SAM2_AVAILABLE:
     warnings.warn("SAM2 not available. Please install sam2 package.")
 
 
@@ -30,15 +37,25 @@ class SAM2VideoMatting:
         self.device = device
         self.memory_offload = memory_offload
         self.fp16 = fp16
+        self.model_cfg = model_cfg
+        self.checkpoint_path = checkpoint_path
         self.predictor = None
         self.inference_state = None
         self.video_segments = {}
+        self.temp_video_path = None
         
-        self._load_model(model_cfg, checkpoint_path)
+        # Don't load model during init - load lazily when needed
+        self._model_loaded = False
     
     def _load_model(self, model_cfg: str, checkpoint_path: str):
-        """Load SAM2 video predictor with optimizations"""
+        """Load SAM2 video predictor lazily"""
+        if self._model_loaded and self.predictor is not None:
+            return  # Already loaded and predictor exists
+            
         try:
+            # Import heavy SAM2 modules only when needed
+            from sam2.build_sam import build_sam2_video_predictor
+            
             # Check for checkpoint in SAM2 repo structure
             if not Path(checkpoint_path).exists():
                 # Try in segment-anything-2/checkpoints/
@@ -57,36 +74,63 @@ class SAM2VideoMatting:
                         if sam2_repo_path.exists():
                             checkpoint_path = str(sam2_repo_path)
             
-            # Use the config path as-is (should be relative to SAM2 package)
-            # Example: "configs/sam2.1/sam2.1_hiera_l.yaml"
+            print(f"🎯 Loading SAM2 model: {model_cfg}")
+            # Use SAM2's build_sam2_video_predictor which returns the predictor directly
+            # The predictor IS the model - no .model attribute needed
             self.predictor = build_sam2_video_predictor(
-                model_cfg, 
-                checkpoint_path,
+                config_file=model_cfg, 
+                ckpt_path=checkpoint_path,
                 device=self.device
             )
             
-            # Enable memory optimizations
-            if self.memory_offload:
-                self.predictor.fill_hole_area = 8
-            
-            if self.fp16 and self.device == "cuda":
-                self.predictor.model.half()
+            self._model_loaded = True
+            print(f"✅ SAM2 model loaded successfully")
                 
         except Exception as e:
             raise RuntimeError(f"Failed to load SAM2 model: {e}")
     
-    def init_video_state(self, video_frames: List[np.ndarray]) -> None:
+    def init_video_state(self, video_frames: List[np.ndarray] = None, video_path: str = None) -> None:
         """Initialize video inference state"""
-        if self.predictor is None:
-            raise RuntimeError("SAM2 model not loaded")
+        # Load model lazily on first use
+        if not self._model_loaded:
+            self._load_model(self.model_cfg, self.checkpoint_path)
         
-        # Create temporary directory for frames if needed
+        if video_path is not None:
+            # Use video path directly (SAM2's preferred method)
             self.inference_state = self.predictor.init_state(
-            video_path=None,
-            video_frames=video_frames,
+                video_path=video_path,
                 offload_video_to_cpu=self.memory_offload,
                 async_loading_frames=True
             )
+        else:
+            # For frame arrays, we need to save them as a temporary video first
+            
+            if video_frames is None or len(video_frames) == 0:
+                raise ValueError("Either video_path or video_frames must be provided")
+            
+            # Create temporary video file in current directory
+            import uuid
+            temp_video_name = f"temp_sam2_{uuid.uuid4().hex[:8]}.mp4"
+            temp_video_path = Path.cwd() / temp_video_name
+            
+            # Write frames to temporary video
+            height, width = video_frames[0].shape[:2]
+            fourcc = cv2.VideoWriter_fourcc(*'mp4v')
+            writer = cv2.VideoWriter(str(temp_video_path), fourcc, 30.0, (width, height))
+            
+            for frame in video_frames:
+                writer.write(frame)
+            writer.release()
+            
+            # Initialize with temporary video
+            self.inference_state = self.predictor.init_state(
+                video_path=str(temp_video_path),
+                offload_video_to_cpu=self.memory_offload,
+                async_loading_frames=True
+            )
+            
+            # Store temp path for cleanup
+            self.temp_video_path = temp_video_path
     
     def add_person_prompts(self, 
                           frame_idx: int, 
@@ -123,13 +167,16 @@ class SAM2VideoMatting:
         
         return object_ids
     
-    def propagate_masks(self, start_frame: int = 0, max_frames: Optional[int] = None) -> Dict[int, Dict[int, np.ndarray]]:
+    def propagate_masks(self, start_frame: int = 0, max_frames: Optional[int] = None, 
+                       frame_release_interval: int = 50, frame_window_size: int = 30) -> Dict[int, Dict[int, np.ndarray]]:
         """
-        Propagate masks through video
+        Propagate masks through video with Det-SAM2 style memory management
         
         Args:
             start_frame: Starting frame index
             max_frames: Maximum number of frames to process
+            frame_release_interval: Release old frames every N frames
+            frame_window_size: Keep N frames in memory
             
         Returns:
             Dictionary mapping frame_idx -> {obj_id: mask}
@@ -153,9 +200,108 @@ class SAM2VideoMatting:
             
             video_segments[out_frame_idx] = frame_masks
             
-            # Memory management: release old frames periodically
-            if self.memory_offload and out_frame_idx % 100 == 0:
-                self._release_old_frames(out_frame_idx - 50)
+            # Det-SAM2 style memory management: more aggressive frame release
+            if self.memory_offload and out_frame_idx % frame_release_interval == 0:
+                self._release_old_frames(out_frame_idx - frame_window_size)
+                # Optional: Log frame release for monitoring
+                if out_frame_idx % (frame_release_interval * 4) == 0:  # Log every 4x release interval
+                    print(f"Det-SAM2: Released frames before {out_frame_idx - frame_window_size}, keeping {frame_window_size} frames")
+        
+        return video_segments
+    
+    def propagate_masks_with_continuous_correction(self, 
+                                                 detector,
+                                                 temp_video_path: str,
+                                                 start_frame: int = 0, 
+                                                 max_frames: Optional[int] = None,
+                                                 correction_interval: int = 60,
+                                                 frame_release_interval: int = 50,
+                                                 frame_window_size: int = 30) -> Dict[int, Dict[int, np.ndarray]]:
+        """
+        Det-SAM2 style: Propagate masks with continuous prompt correction
+        
+        Args:
+            detector: YOLODetector instance for generating correction prompts
+            temp_video_path: Path to video file for frame access
+            start_frame: Starting frame index
+            max_frames: Maximum number of frames to process
+            correction_interval: Add correction prompts every N frames
+            frame_release_interval: Release old frames every N frames
+            frame_window_size: Keep N frames in memory
+            
+        Returns:
+            Dictionary mapping frame_idx -> {obj_id: mask}
+        """
+        if self.inference_state is None:
+            raise RuntimeError("Video state not initialized")
+        
+        video_segments = {}
+        max_frames = max_frames or 10000  # Default limit
+        
+        # Open video for accessing frames during propagation
+        cap = cv2.VideoCapture(str(temp_video_path))
+        
+        try:
+            for out_frame_idx, out_obj_ids, out_mask_logits in self.predictor.propagate_in_video(
+                self.inference_state, 
+                start_frame_idx=start_frame,
+                max_frame_num_to_track=max_frames,
+                reverse=False
+            ):
+                frame_masks = {}
+                
+                for i, out_obj_id in enumerate(out_obj_ids):
+                    mask = (out_mask_logits[i] > 0.0).cpu().numpy()
+                    frame_masks[out_obj_id] = mask
+                
+                video_segments[out_frame_idx] = frame_masks
+                
+                # Det-SAM2 optimization: Add correction prompts at keyframes
+                if (out_frame_idx % correction_interval == 0 and 
+                    out_frame_idx > start_frame and 
+                    out_frame_idx < max_frames - 1):
+                    
+                    # Read frame for detection
+                    cap.set(cv2.CAP_PROP_POS_FRAMES, out_frame_idx)
+                    ret, correction_frame = cap.read()
+                    
+                    if ret:
+                        # Run detection on this keyframe
+                        detections = detector.detect_persons(correction_frame)
+                        
+                        if detections:
+                            # Convert to prompts and add as corrections
+                            box_prompts, labels = detector.convert_to_sam_prompts(detections)
+                            
+                            # Add correction prompts (SAM2 will propagate backward)
+                            correction_count = 0
+                            try:
+                                for i, (box, label) in enumerate(zip(box_prompts, labels)):
+                                    # Use existing object IDs if available, otherwise create new ones
+                                    obj_id = out_obj_ids[i] if i < len(out_obj_ids) else len(out_obj_ids) + i + 1
+                                    
+                                    self.predictor.add_new_points_or_box(
+                                        inference_state=self.inference_state,
+                                        frame_idx=out_frame_idx,
+                                        obj_id=obj_id,
+                                        box=box,
+                                    )
+                                    correction_count += 1
+                                
+                                print(f"Det-SAM2: Added {correction_count} correction prompts at frame {out_frame_idx}")
+                                
+                            except Exception as e:
+                                warnings.warn(f"Failed to add correction prompt at frame {out_frame_idx}: {e}")
+                
+                # Memory management: More aggressive frame release (Det-SAM2 style)
+                if self.memory_offload and out_frame_idx % frame_release_interval == 0:
+                    self._release_old_frames(out_frame_idx - frame_window_size)
+                    # Optional: Log frame release for monitoring
+                    if out_frame_idx % (frame_release_interval * 4) == 0:  # Log every 4x release interval
+                        print(f"Det-SAM2: Released frames before {out_frame_idx - frame_window_size}, keeping {frame_window_size} frames")
+            
+        finally:
+            cap.release()
         
         return video_segments
     
@@ -231,17 +377,58 @@ class SAM2VideoMatting:
         """Clean up resources"""
         if self.inference_state is not None:
             try:
-                if hasattr(self.predictor, 'cleanup_state'):
+                # Reset SAM2 state first (critical for memory cleanup)
+                if self.predictor is not None and hasattr(self.predictor, 'reset_state'):
+                    self.predictor.reset_state(self.inference_state)
+                
+                # Fallback to cleanup_state if available
+                elif self.predictor is not None and hasattr(self.predictor, 'cleanup_state'):
                     self.predictor.cleanup_state(self.inference_state)
+                
+                # Explicitly delete inference state and video segments
+                del self.inference_state
+                if hasattr(self, 'video_segments') and self.video_segments:
+                    del self.video_segments
+                    self.video_segments = {}
+                
             except Exception as e:
                 warnings.warn(f"Failed to cleanup SAM2 state: {e}")
-            
+            finally:
                 self.inference_state = None
         
+        # Clean up temporary video file
+        if self.temp_video_path is not None:
+            try:
+                if self.temp_video_path.exists():
+                    # Remove the temporary video file
+                    self.temp_video_path.unlink()
+                self.temp_video_path = None
+            except Exception as e:
+                warnings.warn(f"Failed to cleanup temp video: {e}")
+        
         # Clear CUDA cache
         if torch.cuda.is_available():
             torch.cuda.empty_cache()
         
+        # Explicitly delete predictor for fresh creation next time
+        if self.predictor is not None:
+            try:
+                del self.predictor
+            except Exception as e:
+                warnings.warn(f"Failed to delete predictor: {e}")
+            finally:
+                self.predictor = None
+        
+        # Reset model loaded state for fresh reload
+        self._model_loaded = False
+        
+        # Force garbage collection (critical for memory leak prevention)
+        gc.collect()
+    
     def __del__(self):
         """Destructor to ensure cleanup"""
+        try:
             self.cleanup()
+        except Exception:
+            # Ignore errors during Python shutdown
+            pass

vr180_matting/video_processor.py

@@ -7,6 +7,12 @@ import tempfile
 import shutil
 from tqdm import tqdm
 import warnings
+import time
+import subprocess
+import gc
+import psutil
+import os
+import sys
 
 from .config import VR180Config
 from .detector import YOLODetector
@@ -35,8 +41,137 @@ class VideoProcessor:
         self.frame_width = 0
         self.frame_height = 0
         
+        # Processing statistics
+        self.processing_stats = {
+            'start_time': None,
+            'end_time': None,
+            'total_duration': 0,
+            'processing_fps': 0,
+            'chunks_processed': 0,
+            'frames_processed': 0
+        }
+        
         self._initialize_models()
     
+    def _get_process_memory_info(self) -> Dict[str, float]:
+        """Get detailed memory usage for current process and children"""
+        current_process = psutil.Process(os.getpid())
+        
+        # Get memory info for current process
+        memory_info = current_process.memory_info()
+        current_rss = memory_info.rss / 1024**3  # Convert to GB
+        current_vms = memory_info.vms / 1024**3  # Virtual memory
+        
+        # Get memory info for all children
+        children_rss = 0
+        children_vms = 0
+        child_count = 0
+        
+        try:
+            for child in current_process.children(recursive=True):
+                try:
+                    child_memory = child.memory_info()
+                    children_rss += child_memory.rss / 1024**3
+                    children_vms += child_memory.vms / 1024**3
+                    child_count += 1
+                except (psutil.NoSuchProcess, psutil.AccessDenied):
+                    pass
+        except psutil.NoSuchProcess:
+            pass
+        
+        # System memory info
+        system_memory = psutil.virtual_memory()
+        system_total = system_memory.total / 1024**3
+        system_available = system_memory.available / 1024**3
+        system_used = system_memory.used / 1024**3
+        system_percent = system_memory.percent
+        
+        return {
+            'process_rss_gb': current_rss,
+            'process_vms_gb': current_vms,
+            'children_rss_gb': children_rss,
+            'children_vms_gb': children_vms,
+            'total_process_gb': current_rss + children_rss,
+            'child_count': child_count,
+            'system_total_gb': system_total,
+            'system_used_gb': system_used,
+            'system_available_gb': system_available,
+            'system_percent': system_percent
+        }
+    
+    def _print_memory_step(self, step_name: str):
+        """Print memory usage for a specific processing step"""
+        memory_info = self._get_process_memory_info()
+        
+        print(f"\n📊 MEMORY: {step_name}")
+        print(f"   Process RSS: {memory_info['process_rss_gb']:.2f} GB")
+        if memory_info['children_rss_gb'] > 0:
+            print(f"   Children RSS: {memory_info['children_rss_gb']:.2f} GB ({memory_info['child_count']} processes)")
+            print(f"   Total Process: {memory_info['total_process_gb']:.2f} GB")
+        print(f"   System: {memory_info['system_used_gb']:.1f}/{memory_info['system_total_gb']:.1f} GB ({memory_info['system_percent']:.1f}%)")
+        print(f"   Available: {memory_info['system_available_gb']:.1f} GB")
+    
+    def _aggressive_memory_cleanup(self, step_name: str = ""):
+        """Perform aggressive memory cleanup and report before/after"""
+        if step_name:
+            print(f"\n🧹 CLEANUP: Before {step_name}")
+        
+        before_info = self._get_process_memory_info()
+        before_rss = before_info['total_process_gb']
+        
+        # Multiple rounds of garbage collection
+        for i in range(3):
+            gc.collect()
+        
+        # Clear torch cache if available
+        try:
+            import torch
+            if torch.cuda.is_available():
+                torch.cuda.empty_cache()
+                torch.cuda.synchronize()
+        except ImportError:
+            pass
+        
+        # Clear OpenCV internal caches
+        try:
+            # Clear OpenCV video capture cache
+            cv2.setUseOptimized(False)
+            cv2.setUseOptimized(True)
+        except Exception:
+            pass
+        
+        # Clear CuPy caches if available  
+        try:
+            import cupy as cp
+            cp._default_memory_pool.free_all_blocks()
+            cp._default_pinned_memory_pool.free_all_blocks()
+            cp.get_default_memory_pool().free_all_blocks()
+            cp.get_default_pinned_memory_pool().free_all_blocks()
+        except ImportError:
+            pass
+        except Exception as e:
+            print(f"   Warning: Could not clear CuPy cache: {e}")
+        
+        # Force Linux to release memory back to OS
+        if sys.platform == 'linux':
+            try:
+                import ctypes
+                libc = ctypes.CDLL("libc.so.6")
+                libc.malloc_trim(0)
+            except Exception as e:
+                print(f"   Warning: Could not trim memory: {e}")
+        
+        # Brief pause to allow cleanup
+        time.sleep(0.1)
+        
+        after_info = self._get_process_memory_info()
+        after_rss = after_info['total_process_gb']
+        freed_memory = before_rss - after_rss
+        
+        if step_name:
+            print(f"   Before: {before_rss:.2f} GB → After: {after_rss:.2f} GB")
+            print(f"   Freed: {freed_memory:.2f} GB")
+    
     def _initialize_models(self):
         """Initialize YOLO detector and SAM2 model"""
         print("Initializing models...")
@@ -146,6 +281,116 @@ class VideoProcessor:
         print(f"Read {len(frames)} frames")
         return frames
     
+    def read_video_frames_dual_resolution(self, 
+                                         video_path: str, 
+                                         start_frame: int = 0, 
+                                         num_frames: Optional[int] = None,
+                                         scale_factor: float = 0.5) -> Dict[str, List[np.ndarray]]:
+        """
+        Read video frames at both original and scaled resolution for dual-resolution processing
+        
+        Args:
+            video_path: Path to video file
+            start_frame: Starting frame index
+            num_frames: Number of frames to read (None for all)
+            scale_factor: Scaling factor for inference frames
+            
+        Returns:
+            Dictionary with 'original' and 'scaled' frame lists
+        """
+        cap = cv2.VideoCapture(video_path)
+        
+        if not cap.isOpened():
+            raise RuntimeError(f"Could not open video file: {video_path}")
+        
+        # Set starting position
+        cap.set(cv2.CAP_PROP_POS_FRAMES, start_frame)
+        
+        original_frames = []
+        scaled_frames = []
+        frame_count = 0
+        
+        # Progress tracking
+        total_to_read = num_frames if num_frames else self.total_frames - start_frame
+        
+        with tqdm(total=total_to_read, desc="Reading dual-resolution frames") as pbar:
+            while True:
+                ret, frame = cap.read()
+                if not ret:
+                    break
+                
+                # Store original frame
+                original_frames.append(frame.copy())
+                
+                # Create scaled frame for inference
+                if scale_factor != 1.0:
+                    new_width = int(frame.shape[1] * scale_factor)
+                    new_height = int(frame.shape[0] * scale_factor)
+                    scaled_frame = cv2.resize(frame, (new_width, new_height), 
+                                            interpolation=cv2.INTER_AREA)
+                else:
+                    scaled_frame = frame.copy()
+                
+                scaled_frames.append(scaled_frame)
+                frame_count += 1
+                pbar.update(1)
+                
+                if num_frames is not None and frame_count >= num_frames:
+                    break
+        
+        cap.release()
+        
+        print(f"Loaded {len(original_frames)} frames:")
+        print(f"  Original: {original_frames[0].shape} per frame")
+        print(f"  Scaled: {scaled_frames[0].shape} per frame (scale_factor={scale_factor})")
+        
+        return {
+            'original': original_frames,
+            'scaled': scaled_frames
+        }
+    
+    def upscale_mask(self, mask: np.ndarray, target_shape: tuple, method: str = 'cubic') -> np.ndarray:
+        """
+        Upscale a mask from inference resolution to original resolution
+        
+        Args:
+            mask: Low-resolution mask (H, W)
+            target_shape: Target shape (H, W) for upscaling
+            method: Upscaling method ('nearest', 'cubic', 'area')
+            
+        Returns:
+            Upscaled mask at target resolution
+        """
+        if mask.shape[:2] == target_shape[:2]:
+            return mask  # Already correct size
+        
+        # Ensure mask is 2D
+        if mask.ndim == 3:
+            mask = mask.squeeze()
+        
+        # Choose interpolation method
+        if method == 'nearest':
+            interpolation = cv2.INTER_NEAREST  # Crisp edges, good for sharp subjects
+        elif method == 'cubic':
+            interpolation = cv2.INTER_CUBIC    # Smooth edges, good for most content
+        elif method == 'area':
+            interpolation = cv2.INTER_AREA     # Good for downscaling, not upscaling
+        else:
+            interpolation = cv2.INTER_CUBIC    # Default to cubic
+        
+        # Upscale mask
+        upscaled_mask = cv2.resize(
+            mask.astype(np.uint8), 
+            (target_shape[1], target_shape[0]),  # (width, height) for cv2.resize
+            interpolation=interpolation
+        )
+        
+        # Convert back to boolean if it was originally boolean
+        if mask.dtype == bool:
+            upscaled_mask = upscaled_mask.astype(bool)
+        
+        return upscaled_mask
+    
     def calculate_optimal_chunking(self) -> Tuple[int, int]:
         """
         Calculate optimal chunk size and overlap based on memory constraints
@@ -234,6 +479,92 @@ class VideoProcessor:
             
             return matted_frames
     
+    def process_chunk_dual_resolution(self, 
+                                    frame_data: Dict[str, List[np.ndarray]], 
+                                    chunk_idx: int = 0) -> List[np.ndarray]:
+        """
+        Process a chunk using dual-resolution approach: inference at low-res, output at full-res
+        
+        Args:
+            frame_data: Dictionary with 'original' and 'scaled' frame lists
+            chunk_idx: Chunk index for logging
+            
+        Returns:
+            List of matted frames at original resolution
+        """
+        original_frames = frame_data['original']
+        scaled_frames = frame_data['scaled']
+        
+        print(f"Processing chunk {chunk_idx} with dual-resolution ({len(original_frames)} frames)")
+        print(f"  Inference: {scaled_frames[0].shape} → Output: {original_frames[0].shape}")
+        
+        with self.memory_manager.memory_monitor(f"dual-res chunk {chunk_idx}"):
+            # Initialize SAM2 with scaled frames for inference
+            self.sam2_model.init_video_state(scaled_frames)
+            
+            # Detect persons in first scaled frame
+            first_scaled_frame = scaled_frames[0]
+            detections = self.detector.detect_persons(first_scaled_frame)
+            
+            if not detections:
+                warnings.warn(f"No persons detected in chunk {chunk_idx}")
+                return self._create_empty_masks(original_frames)
+            
+            print(f"Detected {len(detections)} persons in first frame (at inference resolution)")
+            
+            # Convert detections to SAM2 prompts (detections are already at scaled resolution)
+            box_prompts, labels = self.detector.convert_to_sam_prompts(detections)
+            
+            # Add prompts to SAM2
+            object_ids = self.sam2_model.add_person_prompts(0, box_prompts, labels)
+            print(f"Added prompts for {len(object_ids)} objects")
+            
+            # Propagate masks through chunk at inference resolution
+            video_segments = self.sam2_model.propagate_masks(
+                start_frame=0, 
+                max_frames=len(scaled_frames)
+            )
+            
+            # Apply upscaled masks to original resolution frames
+            matted_frames = []
+            original_shape = original_frames[0].shape[:2]  # (H, W)
+            
+            for frame_idx, original_frame in enumerate(tqdm(original_frames, desc="Applying upscaled masks")):
+                if frame_idx in video_segments:
+                    frame_masks = video_segments[frame_idx]
+                    
+                    # Get combined mask at inference resolution
+                    combined_mask_scaled = self.sam2_model.get_combined_mask(frame_masks)
+                    
+                    if combined_mask_scaled is not None:
+                        # Upscale mask to original resolution
+                        combined_mask_full = self.upscale_mask(
+                            combined_mask_scaled, 
+                            target_shape=original_shape,
+                            method='cubic'  # Smooth upscaling for masks
+                        )
+                        
+                        # Apply upscaled mask to original resolution frame
+                        matted_frame = self.sam2_model.apply_mask_to_frame(
+                            original_frame, combined_mask_full,
+                            output_format=self.config.output.format,
+                            background_color=self.config.output.background_color
+                        )
+                    else:
+                        # No mask for this frame
+                        matted_frame = self._create_empty_mask_frame(original_frame)
+                else:
+                    # No mask for this frame
+                    matted_frame = self._create_empty_mask_frame(original_frame)
+                
+                matted_frames.append(matted_frame)
+            
+            # Cleanup SAM2 state
+            self.sam2_model.cleanup()
+            
+            print(f"✅ Dual-resolution processing complete: {len(matted_frames)} frames at full resolution")
+            return matted_frames
+    
     def _create_empty_masks(self, frames: List[np.ndarray]) -> List[np.ndarray]:
         """Create empty masks when no persons detected"""
         empty_frames = []
@@ -252,19 +583,213 @@ class VideoProcessor:
             # Green screen background
             return np.full_like(frame, self.config.output.background_color, dtype=np.uint8)
     
+    def merge_chunks_streaming(self, chunk_files: List[Path], output_path: str, 
+                             overlap_frames: int = 0, audio_source: str = None) -> None:
+        """
+        Merge processed chunks using streaming approach (no memory accumulation)
+        
+        Args:
+            chunk_files: List of chunk result files (.npz)
+            output_path: Final output video path
+            overlap_frames: Number of overlapping frames
+            audio_source: Audio source file for final video
+        """
+        if not chunk_files:
+            raise ValueError("No chunk files to merge")
+            
+        print(f"🎬 TRUE Streaming merge: {len(chunk_files)} chunks → {output_path}")
+        
+        # Create temporary directory for frame images
+        import tempfile
+        temp_frames_dir = Path(tempfile.mkdtemp(prefix="merge_frames_"))
+        frame_counter = 0
+        
+        try:
+            print(f"📁 Using temp frames dir: {temp_frames_dir}")
+            
+            # Process each chunk frame-by-frame (true streaming)
+            for i, chunk_file in enumerate(chunk_files):
+                print(f"📼 Processing chunk {i+1}/{len(chunk_files)}: {chunk_file.name}")
+                
+                # Load chunk metadata without loading frames array
+                chunk_data = np.load(str(chunk_file))
+                frames_array = chunk_data['frames']  # This is still mmap'd, not loaded
+                total_frames_in_chunk = frames_array.shape[0]
+                
+                # Determine which frames to skip for overlap
+                start_frame_idx = overlap_frames if i > 0 and overlap_frames > 0 else 0
+                frames_to_process = total_frames_in_chunk - start_frame_idx
+                
+                if start_frame_idx > 0:
+                    print(f"   ✂️  Skipping first {start_frame_idx} overlapping frames")
+                
+                print(f"   🔄 Processing {frames_to_process} frames one-by-one...")
+                
+                # Process frames ONE AT A TIME (true streaming)
+                for frame_idx in range(start_frame_idx, total_frames_in_chunk):
+                    # Load only ONE frame at a time
+                    frame = frames_array[frame_idx]  # Load single frame
+                    
+                    # Save frame directly to disk
+                    frame_path = temp_frames_dir / f"frame_{frame_counter:06d}.jpg"
+                    success = cv2.imwrite(str(frame_path), frame, [cv2.IMWRITE_JPEG_QUALITY, 95])
+                    if not success:
+                        raise RuntimeError(f"Failed to save frame {frame_counter}")
+                    
+                    frame_counter += 1
+                    
+                    # Periodic progress and cleanup
+                    if frame_counter % 100 == 0:
+                        print(f"     💾 Saved {frame_counter} frames...")
+                        gc.collect()  # Periodic cleanup
+                
+                print(f"   ✅ Saved {frames_to_process} frames to disk (total: {frame_counter})")
+                
+                # Close chunk file and cleanup
+                chunk_data.close()
+                del chunk_data, frames_array
+                
+                # Don't delete checkpoint files - they're needed for potential resume
+                # The checkpoint system manages cleanup separately
+                print(f"   📋 Keeping checkpoint file: {chunk_file.name}")
+                
+                # Aggressive cleanup and memory monitoring after each chunk
+                self._aggressive_memory_cleanup(f"After streaming merge chunk {i}")
+                
+                # Memory safety check
+                memory_info = self._get_process_memory_info()
+                if memory_info['total_process_gb'] > 35:  # Warning if approaching 46GB limit
+                    print(f"⚠️  High memory usage: {memory_info['total_process_gb']:.1f}GB - forcing cleanup")
+                    gc.collect()
+                    import torch
+                    if torch.cuda.is_available():
+                        torch.cuda.empty_cache()
+                
+            # Create final video directly from frame images using ffmpeg
+            print(f"📹 Creating final video from {frame_counter} frames...")
+            self._create_video_from_frames(temp_frames_dir, Path(output_path), frame_counter)
+            
+            # Add audio if provided
+            if audio_source:
+                self._add_audio_to_video(output_path, audio_source)
+            
+        except Exception as e:
+            print(f"❌ Streaming merge failed: {e}")
+            raise
+            
+        finally:
+            # Cleanup temporary frames directory
+            try:
+                if temp_frames_dir.exists():
+                    import shutil
+                    shutil.rmtree(temp_frames_dir)
+                    print(f"🗑️  Cleaned up temp frames dir: {temp_frames_dir}")
+            except Exception as e:
+                print(f"⚠️  Could not cleanup temp frames dir: {e}")
+            
+            # Memory cleanup
+            gc.collect()
+            
+        print(f"✅ TRUE Streaming merge complete: {output_path}")
+    
+    def _create_video_from_frames(self, frames_dir: Path, output_path: Path, frame_count: int):
+        """Create video directly from frame images using ffmpeg (memory efficient)"""
+        import subprocess
+        
+        frame_pattern = str(frames_dir / "frame_%06d.jpg")
+        fps = self.video_info['fps'] if hasattr(self, 'video_info') and self.video_info else 30.0
+        
+        print(f"🎬 Creating video with ffmpeg: {frame_count} frames at {fps} fps")
+        
+        # Use GPU encoding if available, fallback to CPU
+        gpu_cmd = [
+            'ffmpeg', '-y',  # -y to overwrite output file
+            '-framerate', str(fps),
+            '-i', frame_pattern,
+            '-c:v', 'h264_nvenc',  # NVIDIA GPU encoder
+            '-preset', 'fast',
+            '-cq', '18',           # Quality for GPU encoding
+            '-pix_fmt', 'yuv420p',
+            str(output_path)
+        ]
+        
+        cpu_cmd = [
+            'ffmpeg', '-y',  # -y to overwrite output file
+            '-framerate', str(fps),
+            '-i', frame_pattern,
+            '-c:v', 'libx264',     # CPU encoder
+            '-preset', 'medium',
+            '-crf', '18',          # Quality for CPU encoding
+            '-pix_fmt', 'yuv420p',
+            str(output_path)
+        ]
+        
+        # Try GPU first
+        print(f"🚀 Trying GPU encoding...")
+        result = subprocess.run(gpu_cmd, capture_output=True, text=True)
+        
+        if result.returncode != 0:
+            print("⚠️  GPU encoding failed, using CPU...")
+            print(f"🔄 CPU encoding...")
+            result = subprocess.run(cpu_cmd, capture_output=True, text=True)
+        else:
+            print("✅ GPU encoding successful!")
+        
+        if result.returncode != 0:
+            print(f"❌ FFmpeg stdout: {result.stdout}")
+            print(f"❌ FFmpeg stderr: {result.stderr}")
+            raise RuntimeError(f"FFmpeg failed with return code {result.returncode}")
+        
+        print(f"✅ Video created successfully: {output_path}")
+    
+    def _add_audio_to_video(self, video_path: str, audio_source: str):
+        """Add audio to video using ffmpeg"""
+        import subprocess
+        import tempfile
+        
+        try:
+            # Create temporary file for output with audio
+            temp_path = Path(video_path).with_suffix('.temp.mp4')
+            
+            cmd = [
+                'ffmpeg', '-y',
+                '-i', str(video_path),    # Input video (no audio)
+                '-i', str(audio_source),  # Input audio source
+                '-c:v', 'copy',           # Copy video without re-encoding
+                '-c:a', 'aac',            # Encode audio as AAC
+                '-map', '0:v:0',          # Map video from first input
+                '-map', '1:a:0',          # Map audio from second input
+                '-shortest',              # Match shortest stream duration
+                str(temp_path)
+            ]
+            
+            print(f"🎵 Adding audio: {audio_source} → {video_path}")
+            result = subprocess.run(cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print(f"⚠️  Audio addition failed: {result.stderr}")
+                # Keep original video without audio
+                return
+            
+            # Replace original with audio version
+            Path(video_path).unlink()
+            temp_path.rename(video_path)
+            print(f"✅ Audio added successfully")
+            
+        except Exception as e:
+            print(f"⚠️  Could not add audio: {e}")
+    
     def merge_overlapping_chunks(self, 
                                 chunk_results: List[List[np.ndarray]], 
                                 overlap_frames: int) -> List[np.ndarray]:
         """
-        Merge overlapping chunks with blending in overlap regions
-        
-        Args:
-            chunk_results: List of chunk results
-            overlap_frames: Number of overlapping frames
-            
-        Returns:
-            Merged frame sequence
+        Legacy merge method - DEPRECATED due to memory accumulation
+        Use merge_chunks_streaming() instead for memory efficiency
         """
+        import warnings
+        warnings.warn("merge_overlapping_chunks() is deprecated due to memory accumulation. Use merge_chunks_streaming()", 
+                     DeprecationWarning, stacklevel=2)
+        
         if len(chunk_results) == 1:
             return chunk_results[0]
         
@@ -348,70 +873,307 @@ class VideoProcessor:
         print(f"Saved {len(frames)} PNG frames to {output_dir}")
     
     def _save_mp4_video(self, frames: List[np.ndarray], output_path: str):
-        """Save frames as MP4 video"""
+        """Save frames as MP4 video with audio preservation"""
         if not frames:
             return
         
-        height, width = frames[0].shape[:2]
+        output_path = Path(output_path)
+        temp_frames_dir = output_path.parent / f"temp_frames_{output_path.stem}"
+        temp_frames_dir.mkdir(exist_ok=True)
         
-        fourcc = cv2.VideoWriter_fourcc(*'mp4v')
-        writer = cv2.VideoWriter(output_path, fourcc, self.fps, (width, height))
-        
-        for frame in tqdm(frames, desc="Writing video"):
+        try:
+            # Save frames as images
+            print("Saving frames as images...")
+            for i, frame in enumerate(tqdm(frames, desc="Saving frames")):
                 if frame.shape[2] == 4:  # Convert RGBA to BGR
                     frame = cv2.cvtColor(frame, cv2.COLOR_RGBA2BGR)
-            writer.write(frame)
                 
-        writer.release()
+                frame_path = temp_frames_dir / f"frame_{i:06d}.jpg"
+                cv2.imwrite(str(frame_path), frame, [cv2.IMWRITE_JPEG_QUALITY, 95])
+            
+            # Create video with ffmpeg
+            self._create_video_with_ffmpeg(temp_frames_dir, output_path, len(frames))
+            
+        finally:
+            # Cleanup temporary frames
+            if temp_frames_dir.exists():
+                shutil.rmtree(temp_frames_dir)
+    
+    def _create_video_with_ffmpeg(self, frames_dir: Path, output_path: Path, frame_count: int):
+        """Create video using ffmpeg with audio preservation"""
+        frame_pattern = str(frames_dir / "frame_%06d.jpg")
+        
+        if self.config.output.preserve_audio:
+            # Create video with audio from input
+            cmd = [
+                'ffmpeg', '-y',
+                '-framerate', str(self.fps),
+                '-i', frame_pattern,
+                '-i', str(self.config.input.video_path),  # Input video for audio
+                '-c:v', 'h264_nvenc',  # Try GPU encoding first
+                '-preset', 'fast',
+                '-cq', '18',
+                '-c:a', 'copy',  # Copy audio without re-encoding
+                '-map', '0:v:0',  # Map video from frames
+                '-map', '1:a:0',  # Map audio from input video
+                '-shortest',     # Match shortest stream duration
+                '-pix_fmt', 'yuv420p',
+                str(output_path)
+            ]
+        else:
+            # Create video without audio
+            cmd = [
+                'ffmpeg', '-y',
+                '-framerate', str(self.fps),
+                '-i', frame_pattern,
+                '-c:v', 'h264_nvenc',
+                '-preset', 'fast',
+                '-cq', '18',
+                '-pix_fmt', 'yuv420p',
+                str(output_path)
+            ]
+        
+        print(f"Creating video with ffmpeg...")
+        result = subprocess.run(cmd, capture_output=True, text=True)
+        
+        if result.returncode != 0:
+            # Try CPU encoding as fallback
+            print("GPU encoding failed, trying CPU encoding...")
+            cmd[cmd.index('h264_nvenc')] = 'libx264'
+            cmd[cmd.index('-cq')] = '-crf'  # Change quality parameter for CPU
+            
+            result = subprocess.run(cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print(f"FFmpeg stdout: {result.stdout}")
+                print(f"FFmpeg stderr: {result.stderr}")
+                raise RuntimeError(f"FFmpeg failed with return code {result.returncode}")
+        
+        # Verify frame count if sync verification is enabled
+        if self.config.output.verify_sync:
+            self._verify_frame_count(output_path, frame_count)
+        
         print(f"Saved video to {output_path}")
     
+    def _verify_frame_count(self, video_path: Path, expected_frames: int):
+        """Verify output video has correct frame count"""
+        try:
+            probe = ffmpeg.probe(str(video_path))
+            video_stream = next(
+                (stream for stream in probe['streams'] if stream['codec_type'] == 'video'), 
+                None
+            )
+            
+            if video_stream:
+                actual_frames = int(video_stream.get('nb_frames', 0))
+                if actual_frames != expected_frames:
+                    print(f"⚠️  Frame count mismatch: expected {expected_frames}, got {actual_frames}")
+                else:
+                    print(f"✅ Frame count verified: {actual_frames} frames")
+        except Exception as e:
+            print(f"⚠️  Could not verify frame count: {e}")
+    
     def process_video(self) -> None:
-        """Main video processing pipeline"""
+        """Main video processing pipeline with checkpoint/resume support"""
+        self.processing_stats['start_time'] = time.time()
         print("Starting VR180 video processing...")
         
         # Load video info
         self.load_video_info(self.config.input.video_path)
         
+        # Initialize checkpoint manager
+        from .checkpoint_manager import CheckpointManager
+        checkpoint_mgr = CheckpointManager(
+            self.config.input.video_path,
+            self.config.output.path
+        )
+        
+        # Check for existing checkpoints
+        resume_info = checkpoint_mgr.get_resume_info()
+        if resume_info['can_resume']:
+            print(f"\n🔄 RESUME DETECTED:")
+            print(f"   Found {resume_info['completed_chunks']} completed chunks")
+            print(f"   Continue from where we left off? (saves time!)")
+            checkpoint_mgr.print_status()
+        
         # Calculate chunking parameters
         chunk_size, overlap_frames = self.calculate_optimal_chunking()
         
-        # Process video in chunks
-        chunk_results = []
+        # Calculate total chunks
+        total_chunks = 0
+        for _ in range(0, self.total_frames, chunk_size - overlap_frames):
+            total_chunks += 1
+        checkpoint_mgr.set_total_chunks(total_chunks)
         
+        # Process video in chunks
+        chunk_files = []  # Store file paths instead of frame data
+        temp_chunk_dir = Path(tempfile.mkdtemp(prefix="vr180_chunks_"))
+        
+        try:
+            chunk_idx = 0
             for start_frame in range(0, self.total_frames, chunk_size - overlap_frames):
                 end_frame = min(start_frame + chunk_size, self.total_frames)
                 frames_to_read = end_frame - start_frame
                 
-            chunk_idx = len(chunk_results)
+                # Check if this chunk was already processed
+                existing_chunk = checkpoint_mgr.get_chunk_file(chunk_idx)
+                if existing_chunk:
+                    print(f"\n✅ Chunk {chunk_idx} already processed: {existing_chunk.name}")
+                    chunk_files.append(existing_chunk)
+                    chunk_idx += 1
+                    continue
+                
                 print(f"\nProcessing chunk {chunk_idx}: frames {start_frame}-{end_frame}")
                 
-            # Read chunk frames
+                # Choose processing approach based on scale factor
+                if self.config.processing.scale_factor == 1.0:
+                    # No scaling needed - use original single-resolution approach
+                    print(f"🔄 Reading frames at original resolution (no scaling)")
                     frames = self.read_video_frames(
+                        self.config.input.video_path,
+                        start_frame=start_frame,
+                        num_frames=frames_to_read,
+                        scale_factor=1.0
+                    )
+                    
+                    # Process chunk normally (single resolution)
+                    matted_frames = self.process_chunk(frames, chunk_idx)
+                else:
+                    # Scaling required - use dual-resolution approach
+                    print(f"🔄 Reading frames at dual resolution (scale_factor={self.config.processing.scale_factor})")
+                    frame_data = self.read_video_frames_dual_resolution(
                         self.config.input.video_path,
                         start_frame=start_frame,
                         num_frames=frames_to_read,
                         scale_factor=self.config.processing.scale_factor
                     )
                     
-            # Process chunk
-            matted_frames = self.process_chunk(frames, chunk_idx)
-            chunk_results.append(matted_frames)
+                    # Process chunk with dual-resolution approach
+                    matted_frames = self.process_chunk_dual_resolution(frame_data, chunk_idx)
                 
-            # Memory cleanup
+                # Save chunk to disk immediately to free memory
+                chunk_path = temp_chunk_dir / f"chunk_{chunk_idx:04d}.npz"
+                print(f"Saving chunk {chunk_idx} to disk...")
+                np.savez_compressed(str(chunk_path), frames=matted_frames)
+                
+                # Save to checkpoint
+                checkpoint_mgr.save_chunk(chunk_idx, None, source_chunk_path=chunk_path)
+                
+                chunk_files.append(chunk_path)
+                chunk_idx += 1
+                
+                # Free the frames from memory immediately
+                del matted_frames
+                if self.config.processing.scale_factor == 1.0:
+                    del frames
+                else:
+                    del frame_data
+                
+                # Update statistics
+                self.processing_stats['chunks_processed'] += 1
+                self.processing_stats['frames_processed'] += frames_to_read
+                
+                # Aggressive memory cleanup after each chunk
+                self._aggressive_memory_cleanup(f"chunk {chunk_idx} completion")
+                
+                # Also use memory manager cleanup
                 self.memory_manager.cleanup_memory()
                 
                 if self.memory_manager.should_emergency_cleanup():
                     self.memory_manager.emergency_cleanup()
             
-        # Merge chunks if multiple
-        print("\nMerging chunks...")
-        final_frames = self.merge_overlapping_chunks(chunk_results, overlap_frames)
+            # Mark chunk processing as complete
+            checkpoint_mgr.mark_processing_complete()
             
-        # Save results
-        print(f"Saving {len(final_frames)} processed frames...")
-        self.save_video(final_frames, self.config.output.path)
+            # Check if merge was already done
+            if resume_info.get('merge_complete', False):
+                print("\n✅ Merge already completed in previous run!")
+                print(f"   Output: {self.config.output.path}")
+            else:
+                # Use streaming merge to avoid memory accumulation (fixes OOM)
+                print("\n🎬 Using streaming merge (no memory accumulation)...")
+                
+                # For resume scenarios, make sure we have all chunk files
+                if resume_info['can_resume']:
+                    checkpoint_chunk_files = checkpoint_mgr.get_completed_chunk_files()
+                    if len(checkpoint_chunk_files) != len(chunk_files):
+                        print(f"⚠️  Using {len(checkpoint_chunk_files)} checkpoint files instead of {len(chunk_files)} temp files")
+                        chunk_files = checkpoint_chunk_files
+                
+                # Determine audio source for final video
+                audio_source = None
+                if self.config.output.preserve_audio and Path(self.config.input.video_path).exists():
+                    audio_source = self.config.input.video_path
+                
+                # Stream merge chunks directly to output (no memory accumulation)
+                self.merge_chunks_streaming(
+                    chunk_files=chunk_files,
+                    output_path=self.config.output.path,
+                    overlap_frames=overlap_frames,
+                    audio_source=audio_source
+                )
+                
+                # Mark merge as complete
+                checkpoint_mgr.mark_merge_complete()
+            
+            print("✅ Streaming merge complete - no memory accumulation!")
+            
+            # Calculate final statistics
+            self.processing_stats['end_time'] = time.time()
+            self.processing_stats['total_duration'] = self.processing_stats['end_time'] - self.processing_stats['start_time']
+            if self.processing_stats['total_duration'] > 0:
+                self.processing_stats['processing_fps'] = self.processing_stats['frames_processed'] / self.processing_stats['total_duration']
+            
+            # Print processing statistics
+            self._print_processing_statistics()
             
             # Print final memory report
             self.memory_manager.print_memory_report()
             
             print("Video processing completed!")
+            
+            # Option to clean up checkpoints
+            print("\n🗄️  CHECKPOINT CLEANUP OPTIONS:")
+            print("   Checkpoints saved successfully and can be cleaned up")
+            print("   - Keep checkpoints for debugging: checkpoint_mgr.cleanup_checkpoints(keep_chunks=True)")
+            print("   - Remove all checkpoints: checkpoint_mgr.cleanup_checkpoints()")
+            print(f"   - Checkpoint location: {checkpoint_mgr.checkpoint_dir}")
+            
+            # For now, keep checkpoints by default (user can manually clean)
+            print("\n💡 Checkpoints kept for safety. Delete manually when no longer needed.")
+            
+        finally:
+            # Clean up temporary chunk files (but not checkpoints)
+            if temp_chunk_dir.exists():
+                print("Cleaning up temporary chunk files...")
+                try:
+                    shutil.rmtree(temp_chunk_dir)
+                except Exception as e:
+                    print(f"⚠️  Could not clean temp directory: {e}")
+    
+    def _print_processing_statistics(self):
+        """Print detailed processing statistics"""
+        stats = self.processing_stats
+        video_duration = self.total_frames / self.fps if self.fps > 0 else 0
+        
+        print("\n" + "="*60)
+        print("PROCESSING STATISTICS")
+        print("="*60)
+        print(f"Input video duration: {video_duration:.1f} seconds ({self.total_frames} frames @ {self.fps:.2f} fps)")
+        print(f"Total processing time: {stats['total_duration']:.1f} seconds")
+        print(f"Processing speed: {stats['processing_fps']:.2f} fps")
+        print(f"Speedup factor: {self.fps / stats['processing_fps']:.1f}x slower than realtime")
+        print(f"Chunks processed: {stats['chunks_processed']}")
+        print(f"Frames processed: {stats['frames_processed']}")
+        
+        if video_duration > 0:
+            efficiency = video_duration / stats['total_duration']
+            print(f"Processing efficiency: {efficiency:.3f} (1.0 = realtime)")
+            
+            # Estimate time for different video lengths
+            print(f"\nEstimated processing times:")
+            print(f"  5 minutes: {(5 * 60) / efficiency / 60:.1f} minutes")
+            print(f"  30 minutes: {(30 * 60) / efficiency / 60:.1f} minutes") 
+            print(f"  1 hour: {(60 * 60) / efficiency / 60:.1f} minutes")
+        
+        print("="*60 + "\n")

vr180_matting/vr180_processor.py

@@ -3,6 +3,7 @@ import numpy as np
 from typing import List, Dict, Any, Optional, Tuple
 from pathlib import Path
 import warnings
+import torch
 
 from .video_processor import VideoProcessor
 from .config import VR180Config
@@ -65,17 +66,31 @@ class VR180Processor(VideoProcessor):
         Returns:
             Tuple of (left_eye_frame, right_eye_frame)
         """
-        if self.sbs_split_point == 0:
-            self.sbs_split_point = frame.shape[1] // 2
+        # Always calculate split point based on current frame width
+        # This handles scaled frames correctly
+        frame_width = frame.shape[1]
+        current_split_point = frame_width // 2
         
-        left_eye = frame[:, :self.sbs_split_point]
-        right_eye = frame[:, self.sbs_split_point:]
+        # Debug info on first use
+        if self.sbs_split_point == 0:
+            print(f"Frame dimensions: {frame.shape[1]}x{frame.shape[0]}")
+            print(f"Split point: {current_split_point}")
+            self.sbs_split_point = current_split_point  # Store for reference
+        
+        left_eye = frame[:, :current_split_point]
+        right_eye = frame[:, current_split_point:]
+        
+        # Validate both eyes have content
+        if left_eye.size == 0:
+            raise RuntimeError(f"Left eye frame is empty after split (frame width: {frame_width})")
+        if right_eye.size == 0:
+            raise RuntimeError(f"Right eye frame is empty after split (frame width: {frame_width})")
         
         return left_eye, right_eye
     
     def combine_sbs_frame(self, left_eye: np.ndarray, right_eye: np.ndarray) -> np.ndarray:
         """
-        Combine left and right eye frames back into side-by-side format
+        Combine left and right eye frames back into side-by-side format with GPU acceleration
         
         Args:
             left_eye: Left eye frame
@@ -84,14 +99,44 @@ class VR180Processor(VideoProcessor):
         Returns:
             Combined SBS frame
         """
+        try:
+            import cupy as cp
+            
+            # Transfer to GPU for faster combination
+            left_gpu = cp.asarray(left_eye)
+            right_gpu = cp.asarray(right_eye)
+            
+            # Ensure frames have same height
+            if left_gpu.shape[0] != right_gpu.shape[0]:
+                target_height = min(left_gpu.shape[0], right_gpu.shape[0])
+                # Note: OpenCV resize not available in CuPy, fall back to CPU for resize
+                left_eye = cv2.resize(left_eye, (left_eye.shape[1], target_height))
+                right_eye = cv2.resize(right_eye, (right_eye.shape[1], target_height))
+                left_gpu = cp.asarray(left_eye)
+                right_gpu = cp.asarray(right_eye)
+            
+            # Combine horizontally on GPU (much faster for large arrays)
+            combined_gpu = cp.hstack([left_gpu, right_gpu])
+            
+            # Transfer back to CPU and ensure we get a copy, not a view
+            combined = cp.asnumpy(combined_gpu).copy()
+            
+            # Free GPU memory immediately
+            del left_gpu, right_gpu, combined_gpu
+            cp._default_memory_pool.free_all_blocks()
+            
+            return combined
+            
+        except ImportError:
+            # Fallback to CPU NumPy
             # Ensure frames have same height
             if left_eye.shape[0] != right_eye.shape[0]:
                 target_height = min(left_eye.shape[0], right_eye.shape[0])
                 left_eye = cv2.resize(left_eye, (left_eye.shape[1], target_height))
                 right_eye = cv2.resize(right_eye, (right_eye.shape[1], target_height))
             
-        # Combine horizontally
-        combined = np.hstack([left_eye, right_eye])
+            # Combine horizontally and ensure we get a copy, not a view
+            combined = np.hstack([left_eye, right_eye]).copy()
             return combined
     
     def process_with_disparity_mapping(self, 
@@ -113,8 +158,23 @@ class VR180Processor(VideoProcessor):
         left_eye_frames = []
         right_eye_frames = []
         
-        for frame in frames:
+        for i, frame in enumerate(frames):
             left, right = self.split_sbs_frame(frame)
+            
+            # Debug: Check if frames are valid
+            if i == 0:  # Only debug first frame
+                print(f"Original frame shape: {frame.shape}")
+                print(f"Left eye shape: {left.shape}")
+                print(f"Right eye shape: {right.shape}")
+                print(f"Left eye min/max: {left.min()}/{left.max()}")
+                print(f"Right eye min/max: {right.min()}/{right.max()}")
+            
+            # Validate frames
+            if left.size == 0:
+                raise RuntimeError(f"Left eye frame {i} is empty")
+            if right.size == 0:
+                raise RuntimeError(f"Right eye frame {i} is empty")
+            
             left_eye_frames.append(left)
             right_eye_frames.append(right)
         
@@ -123,6 +183,10 @@ class VR180Processor(VideoProcessor):
         with self.memory_manager.memory_monitor(f"left eye chunk {chunk_idx}"):
             left_matted = self._process_eye_sequence(left_eye_frames, "left", chunk_idx)
         
+        # Free left eye frames after processing (before right eye to save memory)
+        del left_eye_frames
+        self._aggressive_memory_cleanup(f"After left eye processing chunk {chunk_idx}")
+        
         # Process right eye with cross-validation
         print("Processing right eye with cross-validation...")
         with self.memory_manager.memory_monitor(f"right eye chunk {chunk_idx}"):
@@ -130,6 +194,10 @@ class VR180Processor(VideoProcessor):
                 right_eye_frames, left_matted, "right", chunk_idx
             )
         
+        # Free right eye frames after processing
+        del right_eye_frames
+        self._aggressive_memory_cleanup(f"After right eye processing chunk {chunk_idx}")
+        
         # Combine results back to SBS format
         combined_frames = []
         for left_frame, right_frame in zip(left_matted, right_matted):
@@ -140,6 +208,15 @@ class VR180Processor(VideoProcessor):
                 combined = {'left': left_frame, 'right': right_frame}
             combined_frames.append(combined)
         
+        # Free the individual eye results after combining
+        del left_matted
+        del right_matted
+        self._aggressive_memory_cleanup(f"After combining frames chunk {chunk_idx}")
+        
+        # CRITICAL: Complete inter-chunk cleanup to prevent model persistence
+        # This ensures models don't accumulate between chunks
+        self._complete_inter_chunk_cleanup(chunk_idx)
+        
         return combined_frames
     
     def _process_eye_sequence(self, 
@@ -150,16 +227,148 @@ class VR180Processor(VideoProcessor):
         if not eye_frames:
             return []
         
-        # Initialize SAM2 with eye frames
-        self.sam2_model.init_video_state(eye_frames)
+        # Create a unique temporary video for this eye processing
+        import uuid
+        temp_video_name = f"temp_sam2_{eye_name}_chunk{chunk_idx}_{uuid.uuid4().hex[:8]}.mp4"
+        temp_video_path = Path.cwd() / temp_video_name
+        
+        try:
+            # Use ffmpeg approach since OpenCV video writer is failing
+            height, width = eye_frames[0].shape[:2]
+            temp_video_path = temp_video_path.with_suffix('.mp4')
+            
+            print(f"Creating temp video using ffmpeg: {temp_video_path}")
+            print(f"Video params: size=({width}, {height}), frames={len(eye_frames)}")
+            
+            # Create a temporary directory for frame images
+            temp_frames_dir = temp_video_path.parent / f"frames_{temp_video_path.stem}"
+            temp_frames_dir.mkdir(exist_ok=True)
+            
+            # Save frames as individual images (using JPEG for smaller file size)
+            print("Saving frames as images...")
+            for i, frame in enumerate(eye_frames):
+                # Check if frame is empty
+                if frame.size == 0:
+                    raise RuntimeError(f"Frame {i} is empty (size=0)")
+                
+                # Ensure frame is uint8
+                if frame.dtype != np.uint8:
+                    frame = frame.astype(np.uint8)
+                
+                # Debug first frame
+                if i == 0:
+                    print(f"First frame to save: shape={frame.shape}, dtype={frame.dtype}, empty={frame.size == 0}")
+                
+                # Use JPEG instead of PNG for smaller files (faster I/O, less disk space)
+                frame_path = temp_frames_dir / f"frame_{i:06d}.jpg"
+                # Use high quality JPEG to minimize compression artifacts
+                success = cv2.imwrite(str(frame_path), frame, [cv2.IMWRITE_JPEG_QUALITY, 95])
+                if not success:
+                    print(f"Frame {i} details: shape={frame.shape}, dtype={frame.dtype}, size={frame.size}")
+                    raise RuntimeError(f"Failed to save frame {i} as image")
+                
+                if i % 50 == 0:
+                    print(f"Saved {i}/{len(eye_frames)} frames")
+                
+                # Force garbage collection every 100 frames to free memory
+                if i % 100 == 0:
+                    import gc
+                    gc.collect()
+            
+            # Use ffmpeg to create video from images
+            import subprocess
+            # Use the original video's framerate - access through parent class
+            original_fps = self.fps if hasattr(self, 'fps') else 30.0
+            print(f"Using framerate: {original_fps} fps")
+            
+            # Memory monitoring before ffmpeg
+            self._print_memory_step(f"Before ffmpeg encoding ({eye_name} eye)")
+            # Try GPU encoding first, fallback to CPU
+            gpu_cmd = [
+                'ffmpeg', '-y',  # -y to overwrite output file
+                '-framerate', str(original_fps),
+                '-i', str(temp_frames_dir / 'frame_%06d.jpg'),
+                '-c:v', 'h264_nvenc',  # NVIDIA GPU encoder
+                '-preset', 'fast',     # GPU preset
+                '-cq', '18',           # Quality for GPU encoding
+                '-pix_fmt', 'yuv420p',
+                str(temp_video_path)
+            ]
+            
+            cpu_cmd = [
+                'ffmpeg', '-y',  # -y to overwrite output file
+                '-framerate', str(original_fps),
+                '-i', str(temp_frames_dir / 'frame_%06d.jpg'),
+                '-c:v', 'libx264',     # CPU encoder
+                '-pix_fmt', 'yuv420p',
+                '-crf', '18',          # Quality for CPU encoding
+                '-preset', 'medium',
+                str(temp_video_path)
+            ]
+            
+            # Try GPU first
+            print(f"Trying GPU encoding: {' '.join(gpu_cmd)}")
+            result = subprocess.run(gpu_cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print("GPU encoding failed, trying CPU...")
+                print(f"GPU error: {result.stderr}")
+                ffmpeg_cmd = cpu_cmd
+                print(f"Using CPU encoding: {' '.join(ffmpeg_cmd)}")
+                result = subprocess.run(ffmpeg_cmd, capture_output=True, text=True)
+            else:
+                print("GPU encoding successful!")
+                ffmpeg_cmd = gpu_cmd
+            
+            print(f"Running ffmpeg: {' '.join(ffmpeg_cmd)}")
+            result = subprocess.run(ffmpeg_cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print(f"FFmpeg stdout: {result.stdout}")
+                print(f"FFmpeg stderr: {result.stderr}")
+                raise RuntimeError(f"FFmpeg failed with return code {result.returncode}")
+            
+            # Clean up frame images
+            import shutil
+            shutil.rmtree(temp_frames_dir)
+            
+            print(f"Created temp video successfully")
+            
+            # Memory monitoring after ffmpeg
+            self._print_memory_step(f"After ffmpeg encoding ({eye_name} eye)")
+            
+            # Verify the file was created and has content
+            if not temp_video_path.exists():
+                raise RuntimeError(f"Temporary video file was not created: {temp_video_path}")
+            
+            file_size = temp_video_path.stat().st_size
+            if file_size == 0:
+                raise RuntimeError(f"Temporary video file is empty: {temp_video_path}")
+            
+            print(f"Created temp video {temp_video_path} ({file_size / 1024 / 1024:.1f} MB)")
+            
+            # Memory monitoring and cleanup before SAM2 initialization
+            num_frames = len(eye_frames)  # Store count before freeing
+            first_frame = eye_frames[0].copy()  # Copy first frame for detection before freeing
+            self._print_memory_step(f"Before SAM2 init ({eye_name} eye, {num_frames} frames)")
+            
+            # CRITICAL: Explicitly free eye_frames from memory before SAM2 loads the same video
+            # This prevents the OOM issue where both Python frames and SAM2 frames exist simultaneously
+            del eye_frames  # Free the frames array
+            self._aggressive_memory_cleanup(f"SAM2 init for {eye_name} eye")
+            
+            # Initialize SAM2 with video path
+            self._print_memory_step(f"Starting SAM2 init ({eye_name} eye)")
+            self.sam2_model.init_video_state(video_path=str(temp_video_path))
+            self._print_memory_step(f"SAM2 initialized ({eye_name} eye)")
             
             # Detect persons in first frame
-        first_frame = eye_frames[0]
             detections = self.detector.detect_persons(first_frame)
             
             if not detections:
                 warnings.warn(f"No persons detected in {eye_name} eye, chunk {chunk_idx}")
-            return self._create_empty_masks(eye_frames)
+                # Return empty masks for the number of frames
+                return self._create_empty_masks_from_count(num_frames, first_frame.shape)
             
             print(f"Detected {len(detections)} persons in {eye_name} eye first frame")
             
@@ -169,15 +378,45 @@ class VR180Processor(VideoProcessor):
             # Add prompts
             object_ids = self.sam2_model.add_person_prompts(0, box_prompts, labels)
             
-        # Propagate masks
+            # Propagate masks (most expensive operation)
+            self._print_memory_step(f"Before SAM2 propagation ({eye_name} eye, {num_frames} frames)")
+            
+            # Use Det-SAM2 continuous correction if enabled
+            if self.config.matting.continuous_correction:
+                video_segments = self.sam2_model.propagate_masks_with_continuous_correction(
+                    detector=self.detector,
+                    temp_video_path=str(temp_video_path),
+                    start_frame=0, 
+                    max_frames=num_frames,
+                    correction_interval=self.config.matting.correction_interval,
+                    frame_release_interval=self.config.matting.frame_release_interval,
+                    frame_window_size=self.config.matting.frame_window_size
+                )
+                print(f"Used Det-SAM2 continuous correction (interval: {self.config.matting.correction_interval} frames)")
+            else:
                 video_segments = self.sam2_model.propagate_masks(
                     start_frame=0, 
-            max_frames=len(eye_frames)
+                    max_frames=num_frames,
+                    frame_release_interval=self.config.matting.frame_release_interval,
+                    frame_window_size=self.config.matting.frame_window_size
                 )
             
-        # Apply masks
+            self._print_memory_step(f"After SAM2 propagation ({eye_name} eye)")
+            
+            # Apply masks with streaming approach (no frame accumulation)
+            self._print_memory_step(f"Before streaming mask application ({eye_name} eye)")
+            
+            # Process frames one at a time without accumulation
+            cap = cv2.VideoCapture(str(temp_video_path))
             matted_frames = []
-        for frame_idx, frame in enumerate(eye_frames):
+            
+            try:
+                for frame_idx in range(num_frames):
+                    ret, frame = cap.read()
+                    if not ret:
+                        break
+                    
+                    # Apply mask to this single frame
                     if frame_idx in video_segments:
                         frame_masks = video_segments[frame_idx]
                         combined_mask = self.sam2_model.get_combined_mask(frame_masks)
@@ -192,10 +431,34 @@ class VR180Processor(VideoProcessor):
                     
                     matted_frames.append(matted_frame)
                     
-        # Cleanup
+                    # Free the original frame immediately (no accumulation)
+                    del frame
+                    
+                    # Periodic cleanup during processing
+                    if frame_idx % 100 == 0 and frame_idx > 0:
+                        import gc
+                        gc.collect()
+                        
+            finally:
+                cap.release()
+            
+            # Free video segments completely
+            del video_segments  # This holds processed masks from SAM2
+            self._aggressive_memory_cleanup(f"After streaming mask application ({eye_name} eye)")
+            
+            self._print_memory_step(f"Completed streaming mask application ({eye_name} eye)")
+            return matted_frames
+            
+        finally:
+            # Always cleanup
             self.sam2_model.cleanup()
             
-        return matted_frames
+            # Remove temporary video file
+            try:
+                if temp_video_path.exists():
+                    temp_video_path.unlink()
+            except Exception as e:
+                warnings.warn(f"Failed to cleanup temp video {temp_video_path}: {e}")
     
     def _process_eye_sequence_with_validation(self, 
                                             right_eye_frames: List[np.ndarray],
@@ -223,13 +486,17 @@ class VR180Processor(VideoProcessor):
             left_eye_results, right_matted
         )
         
+        # CRITICAL: Free the intermediate results to prevent memory accumulation
+        del left_eye_results  # Don't keep left eye results after validation
+        del right_matted     # Don't keep unvalidated right results
+        
         return validated_results
     
     def _validate_stereo_consistency(self, 
                                    left_results: List[np.ndarray], 
                                    right_results: List[np.ndarray]) -> List[np.ndarray]:
         """
-        Validate and correct stereo consistency between left and right eye results
+        Validate and correct stereo consistency between left and right eye results using GPU acceleration
         
         Args:
             left_results: Left eye processed frames
@@ -238,9 +505,120 @@ class VR180Processor(VideoProcessor):
         Returns:
             Validated right eye frames
         """
+        print(f"🔍 VALIDATION: Starting stereo consistency check ({len(left_results)} frames)")
+        
+        try:
+            import cupy as cp
+            return self._validate_stereo_consistency_gpu(left_results, right_results)
+        except ImportError:
+            print("   Warning: CuPy not available, using CPU validation")
+            return self._validate_stereo_consistency_cpu(left_results, right_results)
+    
+    def _validate_stereo_consistency_gpu(self, 
+                                       left_results: List[np.ndarray], 
+                                       right_results: List[np.ndarray]) -> List[np.ndarray]:
+        """GPU-accelerated batch stereo validation using CuPy with memory-safe batching"""
+        import cupy as cp
+        
+        print("   Using GPU acceleration for stereo validation")
+        
+        # Process in batches to avoid GPU OOM
+        batch_size = 50  # Process 50 frames at a time (safe for 45GB GPU)
+        total_frames = len(left_results)
+        area_ratios_all = []
+        needs_correction_all = []
+        
+        print(f"   Processing {total_frames} frames in batches of {batch_size}...")
+        
+        for batch_start in range(0, total_frames, batch_size):
+            batch_end = min(batch_start + batch_size, total_frames)
+            batch_frames = batch_end - batch_start
+            
+            if batch_start % 100 == 0:
+                print(f"   GPU batch {batch_start//batch_size + 1}: frames {batch_start}-{batch_end}")
+            
+            # Get batch slices
+            left_batch = left_results[batch_start:batch_end]
+            right_batch = right_results[batch_start:batch_end]
+            
+            # Convert batch to GPU
+            left_stack = cp.stack([cp.asarray(frame) for frame in left_batch])
+            right_stack = cp.stack([cp.asarray(frame) for frame in right_batch])
+            
+            # Batch calculate mask areas for this batch
+            if left_stack.shape[3] == 4:  # Alpha channel
+                left_masks = left_stack[:, :, :, 3] > 0
+                right_masks = right_stack[:, :, :, 3] > 0
+            else:  # Green screen detection
+                bg_color = cp.array(self.config.output.background_color)
+                left_diff = cp.abs(left_stack.astype(cp.float32) - bg_color).sum(axis=3)
+                right_diff = cp.abs(right_stack.astype(cp.float32) - bg_color).sum(axis=3)
+                left_masks = left_diff > 30
+                right_masks = right_diff > 30
+            
+            # Calculate areas for this batch
+            left_areas = cp.sum(left_masks, axis=(1, 2))
+            right_areas = cp.sum(right_masks, axis=(1, 2))
+            area_ratios = right_areas.astype(cp.float32) / (left_areas.astype(cp.float32) + 1e-6)
+            
+            # Find frames needing correction in this batch
+            needs_correction = (area_ratios < 0.5) | (area_ratios > 2.0)
+            
+            # Transfer batch results back to CPU and accumulate
+            area_ratios_all.extend(cp.asnumpy(area_ratios))
+            needs_correction_all.extend(cp.asnumpy(needs_correction))
+            
+            # Free GPU memory for this batch
+            del left_stack, right_stack, left_masks, right_masks
+            del left_areas, right_areas, area_ratios, needs_correction
+            cp._default_memory_pool.free_all_blocks()
+        
+        # CRITICAL: Release ALL CuPy memory back to system after validation
+        try:
+            # Force release of all GPU memory pools
+            cp._default_memory_pool.free_all_blocks()
+            cp._default_pinned_memory_pool.free_all_blocks()
+            
+            # Clear CuPy cache completely
+            cp.get_default_memory_pool().free_all_blocks()
+            cp.get_default_pinned_memory_pool().free_all_blocks()
+            
+            print(f"   CuPy memory pools cleared")
+        except Exception as e:
+            print(f"   Warning: Could not clear CuPy memory pools: {e}")
+        
+        correction_count = sum(needs_correction_all)
+        print(f"   GPU validation complete: {correction_count}/{total_frames} frames need correction")
+        
+        # Apply corrections using CPU results
+        validated_frames = []
+        for i, (needs_fix, ratio) in enumerate(zip(needs_correction_all, area_ratios_all)):
+            if i % 100 == 0:
+                print(f"   Processing validation results: {i}/{total_frames}")
+                
+            if needs_fix:
+                # Apply correction
+                corrected_frame = self._apply_stereo_correction(
+                    left_results[i], right_results[i], float(ratio)
+                )
+                validated_frames.append(corrected_frame)
+            else:
+                validated_frames.append(right_results[i])
+        
+        print("✅ VALIDATION: GPU stereo consistency check complete")
+        return validated_frames
+    
+    def _validate_stereo_consistency_cpu(self, 
+                                       left_results: List[np.ndarray], 
+                                       right_results: List[np.ndarray]) -> List[np.ndarray]:
+        """CPU fallback for stereo validation"""
+        print("   Using CPU validation (slower)")
         validated_frames = []
         
         for i, (left_frame, right_frame) in enumerate(zip(left_results, right_results)):
+            if i % 50 == 0:  # Progress every 50 frames
+                print(f"   CPU validation progress: {i}/{len(left_results)}")
+            
             # Simple validation: check if mask areas are similar
             left_mask_area = self._get_mask_area(left_frame)
             right_mask_area = self._get_mask_area(right_frame)
@@ -257,10 +635,44 @@ class VR180Processor(VideoProcessor):
             else:
                 validated_frames.append(right_frame)
         
+        print("✅ VALIDATION: CPU stereo consistency check complete")
         return validated_frames
     
+    def _create_empty_masks_from_count(self, num_frames: int, frame_shape: tuple) -> List[np.ndarray]:
+        """Create empty masks when no persons detected (without frame array)"""
+        empty_frames = []
+        for _ in range(num_frames):
+            if self.config.output.format == "alpha":
+                # Transparent output
+                output = np.zeros((frame_shape[0], frame_shape[1], 4), dtype=np.uint8)
+            else:
+                # Green screen background
+                output = np.full((frame_shape[0], frame_shape[1], 3), 
+                               self.config.output.background_color, dtype=np.uint8)
+            empty_frames.append(output)
+        return empty_frames
+    
     def _get_mask_area(self, frame: np.ndarray) -> float:
-        """Get mask area from processed frame"""
+        """Get mask area from processed frame using GPU acceleration"""
+        try:
+            import cupy as cp
+            
+            # Transfer to GPU
+            frame_gpu = cp.asarray(frame)
+            
+            if frame.shape[2] == 4:  # Alpha channel
+                mask_gpu = frame_gpu[:, :, 3] > 0
+            else:  # Green screen - detect non-background pixels
+                bg_color_gpu = cp.array(self.config.output.background_color)
+                diff_gpu = cp.abs(frame_gpu.astype(cp.float32) - bg_color_gpu).sum(axis=2)
+                mask_gpu = diff_gpu > 30  # Threshold for non-background
+            
+            # Calculate area on GPU and return as Python int
+            area = int(cp.sum(mask_gpu))
+            return area
+            
+        except ImportError:
+            # Fallback to CPU NumPy if CuPy not available
             if frame.shape[2] == 4:  # Alpha channel
                 mask = frame[:, :, 3] > 0
             else:  # Green screen - detect non-background pixels
@@ -284,6 +696,64 @@ class VR180Processor(VideoProcessor):
         # TODO: Implement proper stereo correction algorithm
         return right_frame
     
+    def _complete_inter_chunk_cleanup(self, chunk_idx: int):
+        """
+        Complete inter-chunk cleanup: Destroy all models to prevent memory accumulation
+        
+        This addresses the core issue where SAM2 and YOLO models (~15-20GB) 
+        persist between chunks, causing OOM when processing subsequent chunks.
+        """
+        print(f"🧹 INTER-CHUNK CLEANUP: Destroying all models after chunk {chunk_idx}")
+        
+        # 1. Completely destroy SAM2 model (15-20GB)
+        if hasattr(self, 'sam2_model') and self.sam2_model is not None:
+            self.sam2_model.cleanup()  # Call existing cleanup
+            
+            # Force complete destruction of the model
+            try:
+                # Reset the model's loaded state so it will reload fresh
+                if hasattr(self.sam2_model, '_model_loaded'):
+                    self.sam2_model._model_loaded = False
+                
+                # Clear any cached state
+                if hasattr(self.sam2_model, 'predictor'):
+                    self.sam2_model.predictor = None
+                if hasattr(self.sam2_model, 'inference_state'):
+                    self.sam2_model.inference_state = None
+                
+                print(f"   ✅ SAM2 model destroyed and marked for fresh reload")
+                
+            except Exception as e:
+                print(f"   ⚠️  SAM2 destruction warning: {e}")
+        
+        # 2. Completely destroy YOLO detector (400MB+)
+        if hasattr(self, 'detector') and self.detector is not None:
+            try:
+                # Force YOLO model to be reloaded fresh
+                if hasattr(self.detector, 'model') and self.detector.model is not None:
+                    del self.detector.model
+                    self.detector.model = None
+                    print(f"   ✅ YOLO model destroyed and marked for fresh reload")
+                    
+            except Exception as e:
+                print(f"   ⚠️  YOLO destruction warning: {e}")
+        
+        # 3. Clear CUDA cache aggressively
+        if torch.cuda.is_available():
+            torch.cuda.empty_cache()
+            torch.cuda.synchronize()  # Wait for all operations to complete
+            print(f"   ✅ CUDA cache cleared")
+        
+        # 4. Force garbage collection
+        import gc
+        collected = gc.collect()
+        print(f"   ✅ Garbage collection: {collected} objects freed")
+        
+        # 5. Memory verification
+        self._print_memory_step(f"After complete inter-chunk cleanup (chunk {chunk_idx})")
+        
+        print(f"🎯 RESULT: Models will reload fresh for next chunk (prevents 15-20GB accumulation)")
+    
     def process_chunk(self, 
                      frames: List[np.ndarray], 
                      chunk_idx: int = 0) -> List[np.ndarray]:
@@ -343,6 +813,9 @@ class VR180Processor(VideoProcessor):
                 combined = {'left': left_frame, 'right': right_frame}
             combined_frames.append(combined)
         
+        # CRITICAL: Complete inter-chunk cleanup for independent processing too
+        self._complete_inter_chunk_cleanup(chunk_idx)
+        
         return combined_frames
     
     def save_video(self, frames: List[np.ndarray], output_path: str):