Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Trackers

All trackers share the same Kalman filter and the same detection input, and differ only in how they associate detections to existing tracks. Pick based on your scene and whether you have a Re-ID model.

TrackerAppearanceMatchingWhen to use
SORTNoneIoUSimple scenes, highest speed, no occlusions
ByteTrackNoneIoU (two-stage)Crowded scenes, low-confidence detections, short occlusions
OC-SORTNoneIoU + velocity (OCM)Frequent brief occlusions, no Re-ID available
DeepSORTRe-ID embeddingsAppearance + IoULong occlusions, dense crowds, identity-sensitive cases
Deep OC-SORTRe-ID embeddingsIoU + velocity + appearanceOcclusions where Re-ID helps recover identities

The Kalman filter uses an 8-dimensional state [x, y, a, h, vx, vy, va, vh], where (x, y) is the box centre, a is the aspect ratio, and h is the height. Detections in [x, y, w, h] (top-left) form are converted to and from this representation internally.