BootsTAP: Bootstrapped Training for Tracking-Any-Point
This work addresses the challenge of tracking points on solid surfaces in real-world videos, which is crucial for enhancing AI's perception of motion, but it is incremental as it builds on existing TAP methods with a self-supervised approach.
The paper tackles the problem of Tracking-Any-Point (TAP) in videos to improve models' understanding of physics and motion, achieving state-of-the-art performance with improvements from 61.3% to 67.4% on TAP-Vid-DAVIS and from 57.2% to 62.5% on TAP-Vid-Kinetics.
To endow models with greater understanding of physics and motion, it is useful to enable them to perceive how solid surfaces move and deform in real scenes. This can be formalized as Tracking-Any-Point (TAP), which requires the algorithm to track any point on solid surfaces in a video, potentially densely in space and time. Large-scale groundtruth training data for TAP is only available in simulation, which currently has a limited variety of objects and motion. In this work, we demonstrate how large-scale, unlabeled, uncurated real-world data can improve a TAP model with minimal architectural changes, using a selfsupervised student-teacher setup. We demonstrate state-of-the-art performance on the TAP-Vid benchmark surpassing previous results by a wide margin: for example, TAP-Vid-DAVIS performance improves from 61.3% to 67.4%, and TAP-Vid-Kinetics from 57.2% to 62.5%. For visualizations, see our project webpage at https://bootstap.github.io/