scott/samyolo_on_segments
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

working still

Browse Source
This commit is contained in:
Scott Register
2025-07-27 14:14:21 -07:00
parent cd7bc54efe
commit 97f12c79a4
2 changed files with 82 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

75
core/sam2_processor.py
View File

@@ -65,7 +65,8 @@ class SAM2Processor:
self.predictor = build_sam2_video_predictor( self.predictor = build_sam2_video_predictor(
config_name, # Use just the config name, not full path config_name, # Use just the config name, not full path
self.checkpoint_path, self.checkpoint_path,
device=device device=device,
overrides=dict(conf=0.95)
) )
# Enable optimizations for CUDA # Enable optimizations for CUDA
@@ -539,13 +540,14 @@ class SAM2Processor:
logger.error(f"Error generating first frame debug masks: {e}") logger.error(f"Error generating first frame debug masks: {e}")
return False return False
def add_multi_frame_prompts_to_predictor(self, inference_state, multi_frame_prompts: Dict[int, List[Dict[str, Any]]]) -> bool: def add_multi_frame_prompts_to_predictor(self, inference_state, multi_frame_prompts: Dict[int, Any]) -> bool:
""" """
Add YOLO detection prompts at multiple frame indices for mid-segment re-detection. Add YOLO prompts at multiple frame indices for mid-segment re-detection.
Supports both bounding box prompts (detection mode) and mask prompts (segmentation mode).
Args: Args:
inference_state: SAM2 inference state inference_state: SAM2 inference state
multi_frame_prompts: Dictionary mapping frame_index -> list of prompt dictionaries multi_frame_prompts: Dictionary mapping frame_index -> prompts (list of dicts for bbox, dict with 'masks' for segmentation)
Returns: Returns:
True if prompts were added successfully True if prompts were added successfully
@@ -554,37 +556,56 @@ class SAM2Processor:
logger.warning("SAM2 Mid-segment: No multi-frame prompts provided") logger.warning("SAM2 Mid-segment: No multi-frame prompts provided")
return False return False
total_prompts = sum(len(prompts) for prompts in multi_frame_prompts.values())
logger.info(f"SAM2 Mid-segment: Adding {total_prompts} prompts across {len(multi_frame_prompts)} frames")
success_count = 0 success_count = 0
total_count = 0 total_count = 0
for frame_idx, prompts in multi_frame_prompts.items(): for frame_idx, prompts_data in multi_frame_prompts.items():
logger.info(f"SAM2 Mid-segment: Processing frame {frame_idx} with {len(prompts)} prompts") # Check if this is segmentation mode (masks) or detection mode (bbox prompts)
if isinstance(prompts_data, dict) and 'masks' in prompts_data:
# Segmentation mode: add masks directly
masks_dict = prompts_data['masks']
logger.info(f"SAM2 Mid-segment: Processing frame {frame_idx} with {len(masks_dict)} YOLO masks")
for i, prompt in enumerate(prompts): for obj_id, mask in masks_dict.items():
obj_id = prompt['obj_id'] total_count += 1
bbox = prompt['bbox'] logger.info(f"SAM2 Mid-segment: Frame {frame_idx}, adding mask for Object {obj_id}")
confidence = prompt.get('confidence', 'unknown')
total_count += 1
logger.info(f"SAM2 Mid-segment: Frame {frame_idx}, Prompt {i+1}/{len(prompts)}: Object {obj_id}, bbox={bbox}, conf={confidence}") try:
self.predictor.add_new_mask(inference_state, frame_idx, obj_id, mask)
logger.info(f"SAM2 Mid-segment: ✓ Frame {frame_idx}, Object {obj_id} mask added successfully")
success_count += 1
try: except Exception as e:
_, out_obj_ids, out_mask_logits = self.predictor.add_new_points_or_box( logger.error(f"SAM2 Mid-segment: ✗ Frame {frame_idx}, Object {obj_id} mask failed: {e}")
inference_state=inference_state, continue
frame_idx=frame_idx, # Key: specify the exact frame index
obj_id=obj_id,
box=bbox.astype(np.float32),
)
logger.info(f"SAM2 Mid-segment: ✓ Frame {frame_idx}, Object {obj_id} added successfully - returned obj_ids: {out_obj_ids}") else:
success_count += 1 # Detection mode: add bounding box prompts (existing logic)
prompts = prompts_data
logger.info(f"SAM2 Mid-segment: Processing frame {frame_idx} with {len(prompts)} bbox prompts")
except Exception as e: for i, prompt in enumerate(prompts):
logger.error(f"SAM2 Mid-segment: ✗ Frame {frame_idx}, Object {obj_id} failed: {e}") obj_id = prompt['obj_id']
continue bbox = prompt['bbox']
confidence = prompt.get('confidence', 'unknown')
total_count += 1
logger.info(f"SAM2 Mid-segment: Frame {frame_idx}, Prompt {i+1}/{len(prompts)}: Object {obj_id}, bbox={bbox}, conf={confidence}")
try:
_, out_obj_ids, out_mask_logits = self.predictor.add_new_points_or_box(
inference_state=inference_state,
frame_idx=frame_idx, # Key: specify the exact frame index
obj_id=obj_id,
box=bbox.astype(np.float32),
)
logger.info(f"SAM2 Mid-segment: ✓ Frame {frame_idx}, Object {obj_id} added successfully - returned obj_ids: {out_obj_ids}")
success_count += 1
except Exception as e:
logger.error(f"SAM2 Mid-segment: ✗ Frame {frame_idx}, Object {obj_id} failed: {e}")
continue
if success_count > 0: if success_count > 0:
logger.info(f"SAM2 Mid-segment: Final result - {success_count}/{total_count} prompts successfully added across {len(multi_frame_prompts)} frames") logger.info(f"SAM2 Mid-segment: Final result - {success_count}/{total_count} prompts successfully added across {len(multi_frame_prompts)} frames")

39
main.py
View File

@@ -507,9 +507,9 @@ def main():
logger.info(f"Pipeline Debug: Using {len(previous_masks)} previous masks for segment {segment_idx}") logger.info(f"Pipeline Debug: Using {len(previous_masks)} previous masks for segment {segment_idx}")
logger.info(f"Pipeline Debug: Previous mask object IDs: {list(previous_masks.keys())}") logger.info(f"Pipeline Debug: Previous mask object IDs: {list(previous_masks.keys())}")
# Handle mid-segment detection if enabled (only when using YOLO prompts, not masks) # Handle mid-segment detection if enabled (works for both detection and segmentation modes)
multi_frame_prompts = None multi_frame_prompts = None
if config.get('advanced.enable_mid_segment_detection', False) and yolo_prompts: if config.get('advanced.enable_mid_segment_detection', False) and (yolo_prompts or has_yolo_masks):
logger.info(f"Mid-segment Detection: Enabled for segment {segment_idx}") logger.info(f"Mid-segment Detection: Enabled for segment {segment_idx}")
# Calculate frame indices for re-detection # Calculate frame indices for re-detection
@@ -538,23 +538,48 @@ def main():
scale=config.get_inference_scale() scale=config.get_inference_scale()
) )
# Convert detections to SAM2 prompts # Convert detections to SAM2 prompts (different handling for segmentation vs detection mode)
multi_frame_prompts = {} multi_frame_prompts = {}
cap = cv2.VideoCapture(segment_info['video_file']) cap = cv2.VideoCapture(segment_info['video_file'])
frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)) frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
cap.release() cap.release()
for frame_idx, detections in multi_frame_detections.items(): for frame_idx, detections in multi_frame_detections.items():
if detections: if detections:
prompts = detector.convert_detections_to_sam2_prompts(detections, frame_width) if has_yolo_masks:
multi_frame_prompts[frame_idx] = prompts # Segmentation mode: convert YOLO masks to SAM2 mask prompts
logger.info(f"Mid-segment Detection: Frame {frame_idx} -> {len(prompts)} SAM2 prompts") frame_masks = {}
for i, detection in enumerate(detections[:2]): # Up to 2 objects
if detection.get('has_mask', False):
mask = detection['mask']
# Resize mask to match inference scale
if config.get_inference_scale() != 1.0:
scale = config.get_inference_scale()
scaled_height = int(frame_height * scale)
scaled_width = int(frame_width * scale)
mask = cv2.resize(mask.astype(np.float32), (scaled_width, scaled_height), interpolation=cv2.INTER_NEAREST)
mask = mask > 0.5
obj_id = i + 1 # Sequential object IDs
frame_masks[obj_id] = mask.astype(bool)
logger.debug(f"Mid-segment Detection: Frame {frame_idx}, Object {obj_id} mask - shape: {mask.shape}, pixels: {np.sum(mask)}")
if frame_masks:
# Store as mask prompts (different format than bbox prompts)
multi_frame_prompts[frame_idx] = {'masks': frame_masks}
logger.info(f"Mid-segment Detection: Frame {frame_idx} -> {len(frame_masks)} YOLO masks")
else:
# Detection mode: convert to bounding box prompts (existing logic)
prompts = detector.convert_detections_to_sam2_prompts(detections, frame_width)
multi_frame_prompts[frame_idx] = prompts
logger.info(f"Mid-segment Detection: Frame {frame_idx} -> {len(prompts)} SAM2 prompts")
logger.info(f"Mid-segment Detection: Generated prompts for {len(multi_frame_prompts)} frames") logger.info(f"Mid-segment Detection: Generated prompts for {len(multi_frame_prompts)} frames")
else: else:
logger.info(f"Mid-segment Detection: No additional frames to process (segment has {total_frames} frames)") logger.info(f"Mid-segment Detection: No additional frames to process (segment has {total_frames} frames)")
elif config.get('advanced.enable_mid_segment_detection', False): elif config.get('advanced.enable_mid_segment_detection', False):
logger.info(f"Mid-segment Detection: Skipped for segment {segment_idx} (no initial YOLO prompts)") logger.info(f"Mid-segment Detection: Skipped for segment {segment_idx} (no initial YOLO data)")
# Process segment with SAM2 # Process segment with SAM2
logger.info(f"Pipeline Debug: Starting SAM2 processing for segment {segment_idx}") logger.info(f"Pipeline Debug: Starting SAM2 processing for segment {segment_idx}")