Learning Pixel Trajectories with Multiscale Contrastive Random Walks
This work provides a unified technique for self-supervised learning in video tasks, addressing a fundamental problem in computer vision, though it is incremental as it builds on existing contrastive random walk formulations.
The paper tackles the challenge of establishing space-time correspondence in video modeling tasks by extending contrastive random walks to pixel-level graphs with a coarse-to-fine hierarchy, achieving performance competitive with strong self-supervised methods for optical flow, keypoint tracking, and video object segmentation.
A range of video modeling tasks, from optical flow to multiple object tracking, share the same fundamental challenge: establishing space-time correspondence. Yet, approaches that dominate each space differ. We take a step towards bridging this gap by extending the recent contrastive random walk formulation to much denser, pixel-level space-time graphs. The main contribution is introducing hierarchy into the search problem by computing the transition matrix between two frames in a coarse-to-fine manner, forming a multiscale contrastive random walk when extended in time. This establishes a unified technique for self-supervised learning of optical flow, keypoint tracking, and video object segmentation. Experiments demonstrate that, for each of these tasks, the unified model achieves performance competitive with strong self-supervised approaches specific to that task. Project webpage: https://jasonbian97.github.io/flowwalk