TAO Long-Tail Results

Download all results for this benchmark

Click on a measure to sort the table accordingly. See below for a more detailed description.


Benchmark Statistics

TrackerHOTAallDetAallAssAallHOTAcomDetAcomAssAcomHOTAuncDetAuncAssAuncHz
Le_Tracker
1. online method 		14.9 12.3 20.1 36.1 30.7 47.8 10.8 8.7 14.6 20.9
STCNTracker
2. 		4.5 5.4 4.6 17.1 19.6 16.7 2.0 2.6 2.2 1.0
A. Athar, J. Luiten, P. Voigtlaender, T. Khurana, A. Dave, B. Leibe, D. Ramanan. BURST: A Benchmark for Unifying Object Recognition, Segmentation and Tracking in Video. In WACV, 2023.
SequencesFramesTrajectoriesBoxes
1521947963167132

Evaluation Measures

Lower is better. Higher is better.
Measure Better Perfect Description
HOTAall higher 100%HOTA all [1]. The HOTA metric evaluated for all classes.
DetAall higher 100%DetA all [1]. Detection Jaccard index averaged over localization thresholds for all classes.
AssAall higher 100%AssA all [1]. Association Jaccard index averaged over all matching detections and then averaged over localization thresholds for all classes.
HOTAcom higher 100%HOTA common [1]. The HOTA metric evaluated for common classes.
DetAcom higher 100%DetA common [1]. Detection Jaccard index averaged over localization thresholds for common classes.
AssAcom higher 100%AssA common [1]. Association Jaccard index averaged over all matching detections and then averaged over localization thresholds for common classes.
HOTAunc higher 100%HOTA uncommon [1]. The HOTA metric evaluated for uncommon classes.
DetAunc higher 100%DetA unccommon [1]. Detection Jaccard index averaged over localization thresholds for uncommon classes.
AssAunc higher 100%AssA uncommon [1]. Association Jaccard index averaged over all matching detections and then averaged over localization thresholds for uncommon classes.
Hz higher Inf.Processing speed (in frames per second excluding the detector) on the benchmark. The frequency is provided by the authors and not officially evaluated by the MOTChallenge.

Legend

Symbol Description
online method This is an online (causal) method, i.e. the solution is immediately available with each incoming frame and cannot be changed at any later time.
using public detections This method used the provided detection set as input.
using private detections This method used a private detection set as input.
new This entry has been submitted or updated less than a week ago.

References:


[1] Jonathon Luiten, A.O. & Leibe, B. HOTA: A Higher Order Metric for Evaluating Multi-Object Tracking. International Journal of Computer Vision, 2020.



Contact