15 Keypoints Is All You Need

TL;DR

This paper introduces KeyTrack, a real-time multi-person pose tracking method that uses only keypoints and a Transformer-based entailment model to improve efficiency and accuracy without relying on RGB or optical flow data.

Contribution

The paper proposes a novel keypoint-only pose tracking approach with a Transformer-based entailment model and a parameter-free keypoint refinement technique, achieving state-of-the-art results efficiently.

Findings

State-of-the-art results on PoseTrack'17 and'18 benchmarks.

Significantly reduced computational requirements compared to existing methods.

Effective keypoint refinement improves pose estimation accuracy.

Abstract

Pose tracking is an important problem that requires identifying unique human pose-instances and matching them temporally across different frames of a video. However, existing pose tracking methods are unable to accurately model temporal relationships and require significant computation, often computing the tracks offline. We present an efficient Multi-person Pose Tracking method, KeyTrack, that only relies on keypoint information without using any RGB or optical flow information to track human keypoints in real-time. Keypoints are tracked using our Pose Entailment method, in which, first, a pair of pose estimates is sampled from different frames in a video and tokenized. Then, a Transformer-based network makes a binary classification as to whether one pose temporally follows another. Furthermore, we improve our top-down pose estimation method with a novel, parameter-free, keypoint refinement technique that improves the keypoint estimates used during the Pose Entailment step. We achieve state-of-the-art results on the PoseTrack'17 and the PoseTrack'18 benchmarks while using only a fraction of the computation required by most other methods for computing the tracking information.