FootFormer: Estimating Stability from Visual Input

TL;DR

FootFormer is a novel cross-modality model that accurately predicts human stability metrics from visual data, outperforming existing methods in estimating foot pressure, contact maps, and CoM, with state-of-the-art results in stability assessment.

Contribution

Introduces FootFormer, a new approach that jointly predicts multiple stability-related measures directly from visual input, improving accuracy over prior methods.

Findings

Achieves statistically significant improvements in foot pressure and contact map estimation.

Outperforms existing methods in stability component prediction.

Sets new state-of-the-art in estimating stability metrics from visual data.

Abstract

We propose FootFormer, a cross-modality approach for jointly predicting human motion dynamics directly from visual input. On multiple datasets, FootFormer achieves statistically significantly better or equivalent estimates of foot pressure distributions, foot contact maps, and center of mass (CoM), as compared with existing methods that generate one or two of those measures. Furthermore, FootFormer achieves SOTA performance in estimating stability-predictive components (CoP, CoM, BoS) used in classic kinesiology metrics. Code and data are available at https://github.com/keatonkraiger/Vision-to-Stability.git.