VI-Net: View-Invariant Quality of Human Movement Assessment

TL;DR

This paper introduces VI-Net, a view-invariant neural network that assesses human movement quality from RGB images without skeleton data, using a two-stage process involving trajectory descriptors and a pre-trained CNN.

Contribution

The paper presents a novel view-invariant approach for movement assessment that does not depend on skeleton data and introduces a new multi-view dataset for rehabilitation movements.

Findings

Achieves 0.66 rank correlation on cross-subject evaluation.

Performs well on unseen views with 0.65 correlation.

Outperforms baseline on single-view dataset KIMORE.

Abstract

We propose a view-invariant method towards the assessment of the quality of human movements which does not rely on skeleton data. Our end-to-end convolutional neural network consists of two stages, where at first a view-invariant trajectory descriptor for each body joint is generated from RGB images, and then the collection of trajectories for all joints are processed by an adapted, pre-trained 2D CNN (e.g. VGG-19 or ResNeXt-50) to learn the relationship amongst the different body parts and deliver a score for the movement quality. We release the only publicly-available, multi-view, non-skeleton, non-mocap, rehabilitation movement dataset (QMAR), and provide results for both cross-subject and cross-view scenarios on this dataset. We show that VI-Net achieves average rank correlation of 0.66 on cross-subject and 0.65 on unseen views when trained on only two views. We also evaluate the proposed method on the single-view rehabilitation dataset KIMORE and obtain 0.66 rank correlation against a baseline of 0.62.