tracklab.visualization package

Submodules

tracklab.visualization.detection module

class tracklab.visualization.detection.DebugDetection(threshold=0.5)[source]

Bases: DetectionVisualizer

Detections are classified by colors:

  • Green is True Positive

  • Yellow is False Positive

  • Red is False Negative

draw_detection(image, detection_pred, detection_gt, metric=None)[source]
class tracklab.visualization.detection.DefaultDetection(print_id=True, print_confidence=False)[source]

Bases: DetectionVisualizer

draw_detection(image, detection_pred, detection_gt, metric=None)[source]
class tracklab.visualization.detection.DetectionStats(print_stats=['state', 'hits', 'age', 'time_since_update', 'matched_with'])[source]

Bases: DetectionVisualizer

draw_detection(image, detection_pred, detection_gt, metric=None)[source]
class tracklab.visualization.detection.EllipseDetection(print_id=True)[source]

Bases: DetectionVisualizer

draw_detection(image, detection_pred, detection_gt, metric=None)[source]
class tracklab.visualization.detection.FullDetection[source]

Bases: DefaultDetection

class tracklab.visualization.detection.SimpleDetectionStats[source]

Bases: DetectionStats

tracklab.visualization.image module

class tracklab.visualization.image.FrameCount[source]

Bases: ImageVisualizer

draw_frame(image, detections_pred, detections_gt, image_pred, image_gt)[source]
class tracklab.visualization.image.IgnoreRegions[source]

Bases: ImageVisualizer

draw_frame(image, detections_pred, detections_gt, image_pred, image_gt)[source]

tracklab.visualization.keypoints module

class tracklab.visualization.keypoints.DefaultKeypoints(threshold=0.4, print_confidence=False)[source]

Bases: DetectionVisualizer

draw_detection(image, detection_pred, detection_gt, metric=None)[source]
class tracklab.visualization.keypoints.FullKeypoints[source]

Bases: DefaultKeypoints

tracklab.visualization.tracking module

class tracklab.visualization.tracking.TrackingLine(max_length: int = 60, vertical_pos: float = 0.0)[source]

Bases: DetectionVisualizer

draw_detection(image, detection_pred, detection_gt, metric=None)[source]
draw_frame(image, detections_pred, detections_gt, image_pred, image_gt)[source]
preproces(video_detections_pred, video_detections_gt, video_image_pred, video_image_gt)[source]

tracklab.visualization.visualization_engine module

class tracklab.visualization.visualization_engine.VisualizationEngine(visualizers: Dict[str, Visualizer], save_images: bool = False, save_videos: bool = False, video_fps: int = 25, process_n_videos: int | None = None, process_n_frames_by_video: int | None = None, **kwargs)[source]

Bases: Callback

Visualization engine from list of visualizers.

Parameters:

  • visualizers – a list of visualizer instances, which must implement draw_frame, or subclass DetectionVisualizer and implement draw_detection.

  • save_images – whether to save the visualization as image files (.jpeg)

  • save_videos – whether to save the visualization as video files (.mp4)

  • process_n_videos – number of videos to visualize. Will visualize the first N videos.

  • process_n_frames_by_video – number of frames per video to visualize. Will visualize frames every N/n frames (not first n frames)

draw_frame(image_metadata, detections_pred, detections_gt, image_pred, image_gt, nframes)[source]
on_dataset_track_end(engine: TrackingEngine)[source]
on_video_loop_end(engine, video_metadata, video_idx, detections, image_pred)[source]
visualize(tracker_state: TrackerState, video_id, detections, image_preds, progress=None)[source]
tracklab.visualization.visualization_engine.create_draw_args(image_id, instance, image_metadatas, detections_pred, detections_gt, image_gts, image_preds, nframes)[source]
tracklab.visualization.visualization_engine.get_group(g, key)[source]
tracklab.visualization.visualization_engine.process_frame(args)[source]

tracklab.visualization.visualizer module

class tracklab.visualization.visualizer.DetectionVisualizer[source]

Bases: Visualizer, ABC

color(detection, is_prediction, color_type='default')[source]
abstract draw_detection(image, detection_pred, detection_gt, metric=None)[source]
draw_frame(image, detections_pred, detections_gt, image_pred, image_gt)[source]
post_init(colors, **kwargs)[source]
class tracklab.visualization.visualizer.ImageVisualizer[source]

Bases: Visualizer, ABC

class tracklab.visualization.visualizer.Visualizer[source]

Bases: ABC

abstract draw_frame(image, detections_pred, detections_gt, image_pred, image_gt)[source]
post_init(**kwargs)[source]
preproces(video_detections_pred, video_detections_gt, video_image_pred, video_image_gt)[source]
tracklab.visualization.visualizer.get_fixed_colors(N)[source]