DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion
Peize Sun* Jinkun Cao* Yi Jiang Zehuan Yuan Song Bai Kris Kitani Ping Luo
The University of Hong Kong Carnegie Mellon University ByteDance Inc.
*: equal contribution
A typical pipeline for multi-object tracking (MOT) is to use a detector for object localization, and following re- identification (re-ID) for object association. This pipeline is partially motivated by recent progress in both object detec- tion and re-ID, and partially motivated by biases in existing tracking datasets, where most objects tend to have distin- guishing appearance and re-ID models are sufficient for es- tablishing associations. In response to such bias, we would like to re-emphasize that methods for multi-object track- ing should also work when object appearance is not suffi- ciently discriminative. To this end, we propose a large-scale dataset for multi-human tracking, where humans have sim- ilar appearance, diverse motion and extreme articulation. As the dataset contains mostly group dancing videos, we name it “DanceTrack”. We expect DanceTrack to provide a better platform to develop more MOT algorithms that rely less on visual discrimination and depend more on motion analysis. We benchmark several state-of-the-art trackers on our dataset and observe a significant performance drop on DanceTrack when compared against existing benchmarks.
Benchmark Results
Paper
Arxiv tech report, 2021
Citation
Peize Sun, Jinkun Cao, Yi Jiang, Zehuan Yuan, Song Bai, Kris Kitani, Pingluo. "DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion", 2021. [Bibtex]
Resources
Github: Toolkit and baselines
Codalab: A public competition
Dataset: [Google Drive] [Baidu Drive (code:awew)]
Acknowledgement
We would like to thank the annotator teams and coordi- nators. We also like to thank Xinshuo Weng, Yifu Zhang for valuable discussion and suggestions, Vivek Roy, Pedro Morgado, Shuyang Sun for proof reading. The layout of this page is developed referring to an amazing GAN webpage.