Saved in:
Bibliographic Details
Main Authors: Chen, Erdong, Ji, Yuyang, Greenberg, Jacob K., Steel, Benjamin, Arkam, Faraz, Lewis, Abigail, Singh, Pranay, Liu, Feng
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.08564
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914380112723968
author Chen, Erdong
Ji, Yuyang
Greenberg, Jacob K.
Steel, Benjamin
Arkam, Faraz
Lewis, Abigail
Singh, Pranay
Liu, Feng
author_facet Chen, Erdong
Ji, Yuyang
Greenberg, Jacob K.
Steel, Benjamin
Arkam, Faraz
Lewis, Abigail
Singh, Pranay
Liu, Feng
contents Video-based Clinical Gait Analysis often suffers from poor generalization as models overfit environmental biases instead of capturing pathological motion. To address this, we propose BioGait-VLM, a tri-modal Vision-Language-Biomechanics framework for interpretable clinical gait assessment. Unlike standard video encoders, our architecture incorporates a Temporal Evidence Distillation branch to capture rhythmic dynamics and a Biomechanical Tokenization branch that projects 3D skeleton sequences into language-aligned semantic tokens. This enables the model to explicitly reason about joint mechanics independent of visual shortcuts. To ensure rigorous benchmarking, we augment the public GAVD dataset with a high-fidelity Degenerative Cervical Myelopathy (DCM) cohort to form a unified 8-class taxonomy, establishing a strict subject-disjoint protocol to prevent data leakage. Under this setting, BioGait-VLM achieves state-of-the-art recognition accuracy. Furthermore, a blinded expert study confirms that biomechanical tokens significantly improve clinical plausibility and evidence grounding, offering a path toward transparent, privacy-enhanced gait assessment.
format Preprint
id arxiv_https___arxiv_org_abs_2603_08564
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle BioGait-VLM: A Tri-Modal Vision-Language-Biomechanics Framework for Interpretable Clinical Gait Assessment
Chen, Erdong
Ji, Yuyang
Greenberg, Jacob K.
Steel, Benjamin
Arkam, Faraz
Lewis, Abigail
Singh, Pranay
Liu, Feng
Computer Vision and Pattern Recognition
Video-based Clinical Gait Analysis often suffers from poor generalization as models overfit environmental biases instead of capturing pathological motion. To address this, we propose BioGait-VLM, a tri-modal Vision-Language-Biomechanics framework for interpretable clinical gait assessment. Unlike standard video encoders, our architecture incorporates a Temporal Evidence Distillation branch to capture rhythmic dynamics and a Biomechanical Tokenization branch that projects 3D skeleton sequences into language-aligned semantic tokens. This enables the model to explicitly reason about joint mechanics independent of visual shortcuts. To ensure rigorous benchmarking, we augment the public GAVD dataset with a high-fidelity Degenerative Cervical Myelopathy (DCM) cohort to form a unified 8-class taxonomy, establishing a strict subject-disjoint protocol to prevent data leakage. Under this setting, BioGait-VLM achieves state-of-the-art recognition accuracy. Furthermore, a blinded expert study confirms that biomechanical tokens significantly improve clinical plausibility and evidence grounding, offering a path toward transparent, privacy-enhanced gait assessment.
title BioGait-VLM: A Tri-Modal Vision-Language-Biomechanics Framework for Interpretable Clinical Gait Assessment
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.08564