Introduction

We propose the first benchmark, named Wild Animal Tracking Benchmark (WATB), to encourage further progress of research and applications of visual object tracking in wildlife protection.
 Our tracking benchmarks include：
 If you want more information about WATB, Please check out the benchmark details at WATB Benchmark page, evaluation toolkit and sample results at Evaluation.

Update Performance

We evaluated the performance of some latest algorithms on WATB, and the performance evaluation results are as follows. Please refer to past algorithm evaluation results in Overall all Performance.

Trackers Description Language AUC Norm Pre Pre
AiATrack ECCV22 Python 0.541 0.616 0.620
TATrack AAAI23 Python 0.525 0.588 0.598
SeqTrack CVPR23 Python 0.507 0.564 0.575
LightFC KBS24 Python 0.501 0.569 0.573
SparseTT IJCAI22 Python 0.499 0.557 0.565
GRM CVPR23 Python 0.491 0.539 0.549
SiamRBO CVPR22 Python 0.464 0.539 0.548
TCTrack++ IEEE23 Python 0.437 0.502 0.509
ETTrack WACV23 Python 0.454 0.512 0.506
CTTrack AAAI23 Python 0.450 0.489 0.499
SGDViT IEEE23 Python 0.427 0.486 0.482
SiamTPN WACV22 Python 0.417 0.471 0.462
precision normalized precision success measures

Video

We selected three labeled video sequences from 206 wild-animal-sequence for display. These three videos represent sea, land and air respectively. These wild-animal-sequence are annotated manually, and then the annotated wild-animal-sequence are reviewed. For more wild-animal-sequence, please see Benchmark page.

Acknowledgements

We refer to OTB data set and TOTB data set in the preliminary work of preparing to establish wildlife tracking benchmark. If you use WATB for research, please consider to cite the following papers:
Copyright @ WATB Team 2022