AnimalTrack: A Benchmark for Multi-Animal Tracking in the Wild

Libo Zhang*        Junyuan Gao*        Zhen Xiao        Heng Fan†
*Equal contribution        †Corresponding author

Data Format

For each video sequence in AnimalTrack, we provide high-accuracy bounding box annotations for all targets. The format of a bounding box is defined as follows
[frame_id, targer_id, top_left_x, top_left_y, width, height, confidence, class, visibility]
Please check out detailed definitions in the paper.

Download the whole dataset

Download the whole AnimalTrack (labels included) with all 58 sequences in the below:
AnimalTrack - google drive OR AnimalTrack - baidu pan

Download the Training/Testing splits in the below:
Training/Testing Splits - google drive OR Training/Testing Splits - baidu pan

Disclaimer: All videos are obtained from the Internet and used for non-commercial research purposes only. We are not responsible for the content or the meaning of these videos. Downloading the videos in AnimalTrack means that you understand and agree the license here.