TrackMAE: Video Representation Learning via Track Mask and Predict

TL;DR

TrackMAE introduces a novel masked video modeling approach that explicitly incorporates motion information through point tracking and motion-aware masking, leading to improved video representations for motion-centric tasks.

Contribution

It proposes using point trajectories and motion-aware masking in masked video modeling to enhance the encoding of temporal dynamics.

Findings

Outperforms state-of-the-art video self-supervised learning methods on six datasets.

Learns more discriminative and generalizable video representations.

Improves performance on motion-centric downstream tasks.

Abstract

Masked video modeling (MVM) has emerged as a simple and scalable self-supervised pretraining paradigm, but only encodes motion information implicitly, limiting the encoding of temporal dynamics in the learned representations. As a result, such models struggle on motion-centric tasks that require fine-grained motion awareness. To address this, we propose TrackMAE, a simple masked video modeling paradigm that explicitly uses motion information as a reconstruction signal. In TrackMAE, we use an off-the-shelf point tracker to sparsely track points in the input videos, generating motion trajectories. Furthermore, we exploit the extracted trajectories to improve random tube masking with a motion-aware masking strategy. We enhance video representations learned in both pixel and feature semantic reconstruction spaces by providing a complementary supervision signal in the form of motion targets. We evaluate on six datasets across diverse downstream settings and find that TrackMAE consistently outperforms state-of-the-art video self-supervised learning baselines, learning more discriminative and generalizable representations. Code available at https://github.com/rvandeghen/TrackMAE