All Keypoints You Need: Detecting Arbitrary Keypoints on the Body of Triple, High, and Long Jump Athletes

TL;DR

This paper introduces a method for detecting arbitrary body keypoints in athletes performing jumps, enhancing performance analysis by capturing detailed body postures beyond standard keypoints.

Contribution

It presents a novel approach leveraging limited annotations and segmentation masks to detect diverse keypoints, including bent joints, on athletes' bodies.

Findings

Effective detection of keypoints on various body parts including bent joints

Model outperforms baseline methods in keypoint localization accuracy

Versatile encoding techniques improve keypoint detection performance

Abstract

Performance analyses based on videos are commonly used by coaches of athletes in various sports disciplines. In individual sports, these analyses mainly comprise the body posture. This paper focuses on the disciplines of triple, high, and long jump, which require fine-grained locations of the athlete's body. Typical human pose estimation datasets provide only a very limited set of keypoints, which is not sufficient in this case. Therefore, we propose a method to detect arbitrary keypoints on the whole body of the athlete by leveraging the limited set of annotated keypoints and auto-generated segmentation masks of body parts. Evaluations show that our model is capable of detecting keypoints on the head, torso, hands, feet, arms, and legs, including also bent elbows and knees. We analyze and compare different techniques to encode desired keypoints as the model's input and their embedding for the Transformer backbone.