Analyzing Data Efficiency and Performance of Machine Learning Algorithms for Assessing Low Back Pain Physical Rehabilitation Exercises

TL;DR

This paper compares the effectiveness of RGB-D and RGB-based human pose estimation methods in assessing low back pain rehabilitation exercises using a Gaussian Mixture Model for performance evaluation.

Contribution

It introduces a comparative analysis of pose estimation techniques for rehabilitation assessment and evaluates their performance on clinical patient data.

Findings

RGB-D and RGB pose estimation methods show comparable assessment accuracy.

GMM-based metrics effectively evaluate patient performance.

The study highlights the potential of RGB-based methods for low-cost rehabilitation monitoring.

Abstract

Analyzing human motion is an active research area, with various applications. In this work, we focus on human motion analysis in the context of physical rehabilitation using a robot coach system. Computer-aided assessment of physical rehabilitation entails evaluation of patient performance in completing prescribed rehabilitation exercises, based on processing movement data captured with a sensory system, such as RGB and RGB-D cameras. As 2D and 3D human pose estimation from RGB images had made impressive improvements, we aim to compare the assessment of physical rehabilitation exercises using movement data obtained from both RGB-D camera (Microsoft Kinect) and estimation from RGB videos (OpenPose and BlazePose algorithms). A Gaussian Mixture Model (GMM) is employed from position (and orientation) features, with performance metrics defined based on the log-likelihood values from GMM. The evaluation is performed on a medical database of clinical patients carrying out low back-pain rehabilitation exercises, previously coached by robot Poppy.