Detector

class byotrack.api.detector.Detector

Bases: ABC

Base class for detections in videos.

Detect on each frame the objects of interest.

abstractmethod run(video: Sequence[np.ndarray] | np.ndarray) → Sequence[byotrack.Detections]

Run the detector on a whole video.

Parameters:: video (Sequence[np.ndarray] | np.ndarray) – Sequence of T frames (array). Each array is expected to have a shape ([D, ]H, W, C)
Returns:: Detections for each frame (ordered by frames)
Return type:: Sequence[byotrack.Detections]

class byotrack.api.detector.BatchDetector(batch_size: int = 20)

Bases: Detector

Abstract detector that performs detection directly by batch.

Usually leads to faster implementation of the detection process when batch size is greater than 1.

batch_size

Size of the frame batch. Default: 20

Type:: int

run(video: Sequence[np.ndarray] | np.ndarray) → list[byotrack.Detections]

Run the detector on a whole video.

Parameters:: video (Sequence[np.ndarray] | np.ndarray) – Sequence of T frames (array). Each array is expected to have a shape ([D, ]H, W, C)
Returns:: Detections for each frame (ordered by frames)
Return type:: Sequence[byotrack.Detections]

abstractmethod detect(batch: np.ndarray) → Sequence[byotrack.Detections]

Apply the detection on a batch of frames.

Parameters:: batch (np.ndarray) – Batch of video frames Shape: (B, [D, ]H, W, C)
Returns:: Detections for each given frame
Return type:: Sequence[byotrack.Detections]

class byotrack.api.detector.DetectionsRefiner

Bases: ABC

Abstract class for method to improve a coarse detection result.

This assumes the refining can be done independently for each frame.

Warning: The class is still experimental and may change in future versions

abstractmethod apply(detections: Detections | ndarray | Tensor, frame: ndarray | None = None) → Detections

Refines Detections of the given frame.

Parameters:

detections (byotrack.DetectionsLike) – Detections to refine
frame (np.ndarray | None) – The associated frame of the video (optional) Shape: ([D, ]H, W, C) Default: None

Returns:

Refined detections

Return type:

byotrack.Detections

run(detections_sequence: Sequence[byotrack.DetectionsLike], video: Sequence[np.ndarray] | np.ndarray | None = None) → list[byotrack.Detections]

Run the refiner on a full video / detections sequence.

Parameters:

detections_sequence (Sequence[byotrack.DetectionsLike]) – Sequence of T detections to refine
video (Sequence[np.ndarray] | np.ndarray | None) – Optional corresponding sequence of T frames. Each frame is expected to have a shape ([D, ]H, W, C)

Returns:

Sequence of the T refined detections

Return type:

Sequence[byotrack.Detections]

class byotrack.api.detector.GroundTruthDetector(segmentations: Sequence[np.ndarray] | np.ndarray | None = None, batch_size=1)

Bases: BatchDetector

Convert a video of segmentations into Detections using the BatchDetector API.

Each frame in the video is expected to be an instance segmentation mask of shape ([D, ]H, W, 1) The video should not be normalized and each pixel is expected to be an integer.

Note: The segmented video can be given at construction time of the detector. In such case,: this video will be used over the one given in run. This allows to use precomputed detections as well as the original video in a [Batch]MultiStepTracker pipeline.

Example:

### Example: Loading CTC segmentation
video = byotrack.Video("dataset/01_ERR_SEG")  # Load segmentation for CLB
# video = byotrack.Video("dataset/01_GT/SEG")  # Load ground-truth segmentation
# video = byotrack.Video("dataset/01_RES/TRA")  # Load predicted tracks segmentation

detector = GroundTruthDetector()
detections_sequence = detector.run(video)

### Example: Loading precomputed segmentations
segmentations = byotrack.Video("segmentations.tiff")  # Shape (T, [D, ]H, W, 1), dtype: int
detections_sequence = GroundTruthDetector().run(segmentations)

### Example: Tracking with precomputed segmentations with online loading
video = byotrack.Video("video.tiff").normalize()  # Shape (T, [D, ]H, W, 1), dtype: float
segmentations = byotrack.Video("segmentations.tiff")  # Shape (T, [D, ]H, W, 1), dtype: int
detector = GroundTruthDetector(segmentations)
linker = ...
tracker = BatchMultiStepTracker(detector, linker)

# This will forward the segmented frame to the Detector,
# but the video frame to the linker.
tracks = tracker.run(video)

run(video: Sequence[np.ndarray] | np.ndarray) → list[byotrack.Detections]

Run the detector on a whole video.

Parameters:: video (Sequence[np.ndarray] | np.ndarray) – Sequence of T frames (array). Each array is expected to have a shape ([D, ]H, W, C)
Returns:: Detections for each frame (ordered by frames)
Return type:: Sequence[byotrack.Detections]

detect(batch: ndarray) → list[SegmentationDetections]

Apply the detection on a batch of frames.

Parameters:: batch (np.ndarray) – Batch of video frames Shape: (B, [D, ]H, W, C)
Returns:: Detections for each given frame
Return type:: Sequence[byotrack.Detections]