Clinical-Prior Guided Multi-Modal Learning with Latent Attention Pooling for Gait-Based Scoliosis Screening

Dong Chen, Zizhuang Wei, Jialei Xu, Xinyang Sun, Zonglin He, Meiru An, Huili Peng, Yong Hu, Kenneth MC Cheung · 2026

Read Clinical-Prior Guided Multi-Modal Learning with Latent Attention Pooling for Gait-Based Scoliosis Screening by Dong Chen, Zizhuang Wei, Jialei Xu, Xinyang Sun, Zonglin He, Meiru An, Huili Peng, Yong Hu, Kenneth MC Cheung (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-06743.txt.

Description

Dong Chen and eight co-authors build a screening tool for adolescent idiopathic scoliosis, a sideways curvature of the spine that early treatment can slow. Ordinary screening is subjective and hard to scale, while video analysis of walking patterns offers a cheaper route. Earlier gait datasets, though, often leaked data by including repeated clips of the same person in training and testing.

Their dataset, ScoliGait, holds 1,572 training clips and 300 fully independent test clips, each labeled with measured spine angles and clinical text. Their model fuses video, text, and a knowledge map built from clinical movement priors, using a latent attention pooling step to combine them. On the leakage-free benchmark it set a new state of the art by a wide margin, and the design keeps its reasoning open to clinical inspection.

Catalog record

CREATORDong Chen, Zizhuang Wei, Jialei Xu, Xinyang Sun, Zonglin He, Meiru An, Huili Peng, Yong Hu, Kenneth MC Cheung
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE18 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-06743

Browse the full library · About the Tower of Babel