Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Luo, Zeqi, Gooya, Ali, Ho, Edmond S. L.
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2507.04814
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915373615415296
author Luo, Zeqi
Gooya, Ali
Ho, Edmond S. L.
author_facet Luo, Zeqi
Gooya, Ali
Ho, Edmond S. L.
contents General movement assessment (GMA) is a non-invasive tool for the early detection of brain dysfunction through the qualitative assessment of general movements, and the development of automated methods can broaden its application. However, mainstream pose-based automated GMA methods are prone to uncertainty due to limited high-quality data and noisy pose estimation, hindering clinical reliability without reliable uncertainty measures. In this work, we introduce UDF-GMA which explicitly models epistemic uncertainty in model parameters and aleatoric uncertainty from data noise for pose-based automated GMA. UDF-GMA effectively disentangles uncertainties by directly modelling aleatoric uncertainty and estimating epistemic uncertainty through Bayesian approximation. We further propose fusing these uncertainties with the embedded motion representation to enhance class separation. Extensive experiments on the Pmi-GMA benchmark dataset demonstrate the effectiveness and generalisability of the proposed approach in predicting poor repertoire.
format Preprint
id arxiv_https___arxiv_org_abs_2507_04814
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle UDF-GMA: Uncertainty Disentanglement and Fusion for General Movement Assessment
Luo, Zeqi
Gooya, Ali
Ho, Edmond S. L.
Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
General movement assessment (GMA) is a non-invasive tool for the early detection of brain dysfunction through the qualitative assessment of general movements, and the development of automated methods can broaden its application. However, mainstream pose-based automated GMA methods are prone to uncertainty due to limited high-quality data and noisy pose estimation, hindering clinical reliability without reliable uncertainty measures. In this work, we introduce UDF-GMA which explicitly models epistemic uncertainty in model parameters and aleatoric uncertainty from data noise for pose-based automated GMA. UDF-GMA effectively disentangles uncertainties by directly modelling aleatoric uncertainty and estimating epistemic uncertainty through Bayesian approximation. We further propose fusing these uncertainties with the embedded motion representation to enhance class separation. Extensive experiments on the Pmi-GMA benchmark dataset demonstrate the effectiveness and generalisability of the proposed approach in predicting poor repertoire.
title UDF-GMA: Uncertainty Disentanglement and Fusion for General Movement Assessment
topic Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
url https://arxiv.org/abs/2507.04814