byte_tracker

A standalone Python implementation of the ByteTrack multi-object tracker based on the official implementation.

BYTETracker

 BYTETracker (track_thresh:float=0.25, track_buffer:int=30,
              match_thresh:float=0.8, frame_rate:int=30)

BYTETracker is a class for tracking objects in video streams using bounding box detections, with methods for processing and updating tracks based on detection results and IoU matching.

	Type	Default	Details
track_thresh	float	0.25	Threshold value for tracking.
track_buffer	int	30	Size of buffer for tracking.
match_thresh	float	0.8	Threshold value for matching tracks to detections.
frame_rate	int	30	Frame rate of the input video stream.

source

BYTETracker.init

 BYTETracker.__init__ (track_thresh:float=0.25, track_buffer:int=30,
                       match_thresh:float=0.8, frame_rate:int=30)

Initializes the BYTETracker.

	Type	Default	Details
track_thresh	float	0.25	Threshold value for tracking.
track_buffer	int	30	Size of buffer for tracking.
match_thresh	float	0.8	Threshold value for matching tracks to detections.
frame_rate	int	30	Frame rate of the input video stream.

source

BYTETracker._process_output

 BYTETracker._process_output (output_results)

Process the output results to separate scores and bounding boxes.

	Type	Details
output_results		Detection results.
Returns	tuple	scores and bounding boxes.

source

BYTETracker._scale_bboxes

 BYTETracker._scale_bboxes (img_info:tuple, img_size:tuple, bboxes:list)

Scale bounding boxes based on image size.

	Type	Details
img_info	tuple	Original height and width of the image.
img_size	tuple	Target size.
bboxes	list	List of bounding boxes.
Returns	list	Scaled bounding boxes.

source

BYTETracker._get_detections

 BYTETracker._get_detections (dets:list, scores_keep:list)

Convert bounding boxes to STrack objects.

	Type	Details
dets	list	List of detections.
scores_keep	list	Scores for the detections.
Returns	list	List of STrack objects.

source

BYTETracker._update_tracked_stracks

 BYTETracker._update_tracked_stracks ()

Update the list of tracked and unconfirmed tracks.

source

BYTETracker._match_tracks_to_detections

 BYTETracker._match_tracks_to_detections (stracks:list, detections:list,
                                          thresh:float)

Match tracks to detections using IOU.

	Type	Details
stracks	list	List of tracks.
detections	list	List of detections.
thresh	float	IOU threshold for matching.
Returns	tuple	Matches and unmatched tracks and detections.

source

BYTETracker._update_tracks

 BYTETracker._update_tracks (stracks:list, detections:list, matches:list,
                             refind_stracks:list, activated_stracks:list)

Update tracks based on matches with detections.

	Type	Details
stracks	list	List of tracks.
detections	list	List of detections.
matches	list	Matched track-detection pairs.
refind_stracks	list	List to add refind tracks.
activated_stracks	list	List to add activated tracks.

source

BYTETracker.update

 BYTETracker.update (output_results, img_info:tuple, img_size:tuple)

Update the tracker based on new detections.

	Type	Details
output_results		Detection results.
img_info	tuple	Original height and width of the image.
img_size	tuple	Target size.
Returns	list	List of activated tracks.

source

joint_stracks

 joint_stracks (track_list_a:list, track_list_b:list)

Combines two lists of tracks ensuring each track is unique based on its track_id.

	Type	Details
track_list_a	list	The first list of tracks.
track_list_b	list	The second list of tracks.
Returns	list	A combined list of unique tracks.

source

sub_stracks

 sub_stracks (track_list_a:list, track_list_b:list)

Subtracts the tracks in track_list_b from track_list_a based on track_id.

	Type	Details
track_list_a	list	The list of tracks to subtract from.
track_list_b	list	The list of tracks to subtract.
Returns	list	A list containing tracks from track_list_a that are not in track_list_b.

source

remove_duplicate_stracks

 remove_duplicate_stracks (s_tracks_a:list, s_tracks_b:list)

Removes duplicate tracks from two lists based on a defined distance metric and time criteria.

	Type	Details
s_tracks_a	list	The first list of tracks.
s_tracks_b	list	The second list of tracks.
Returns	tuple	Two lists of tracks with duplicates removed.