Tracking Emerges by Colorizing Videos
Abstract
We use large amounts of unlabeled video to learn models for visual tracking without manual human supervision. We leverage the natural temporal coherency of color to create a model that learns to colorize gray-scale videos by copying colors from a reference frame. Quantitative and qualitative experiments suggest that this task causes the model to automatically learn to track visual regions. Although the model is trained without any ground-truth labels, our method learns to track well enough to outperform the latest methods based on optical flow. Moreover, our results suggest that failures to track are correlated with failures to colorize, indicating that advancing video colorization may further improve self-supervised visual tracking.
Cite
Text
Vondrick et al. "Tracking Emerges by Colorizing Videos." Proceedings of the European Conference on Computer Vision (ECCV), 2018. doi:10.1007/978-3-030-01261-8_24Markdown
[Vondrick et al. "Tracking Emerges by Colorizing Videos." Proceedings of the European Conference on Computer Vision (ECCV), 2018.](https://mlanthology.org/eccv/2018/vondrick2018eccv-tracking/) doi:10.1007/978-3-030-01261-8_24BibTeX
@inproceedings{vondrick2018eccv-tracking,
title = {{Tracking Emerges by Colorizing Videos}},
author = {Vondrick, Carl and Shrivastava, Abhinav and Fathi, Alireza and Guadarrama, Sergio and Murphy, Kevin},
booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
year = {2018},
doi = {10.1007/978-3-030-01261-8_24},
url = {https://mlanthology.org/eccv/2018/vondrick2018eccv-tracking/}
}