Viewpoint-Invariant Exercise Repetition Counting

TL;DR

This paper introduces a vision-based method for accurately counting exercise repetitions, including concurrent motions, using skeleton data, validated across multiple datasets and camera angles, suitable for remote rehabilitation.

Contribution

The work presents a novel view-angle and motion agnostic approach for counting concurrent exercise motions from skeleton data, applicable in remote settings.

Findings

Mean absolute error (MAE) of 0.06 on datasets

High off-by-one accuracy (OBOA) above 0.88

Effective across various camera angles and concurrent motions

Abstract

Counting the repetition of human exercise and physical rehabilitation is a common task in rehabilitation and exercise training. The existing vision-based repetition counting methods less emphasize the concurrent motions in the same video. This work presents a vision-based human motion repetition counting applicable to counting concurrent motions through the skeleton location extracted from various pose estimation methods. The presented method was validated on the University of Idaho Physical Rehabilitation Movements Data Set (UI-PRMD), and MM-fit dataset. The overall mean absolute error (MAE) for mm-fit was 0.06 with off-by-one Accuracy (OBOA) 0.94. Overall MAE for UI-PRMD dataset was 0.06 with OBOA 0.95. We have also tested the performance in a variety of camera locations and concurrent motions with conveniently collected video with overall MAE 0.06 and OBOA 0.88. The proposed method provides a view-angle and motion agnostic concurrent motion counting. This method can potentially use in large-scale remote rehabilitation and exercise training with only one camera.