Evaluating machine learning muscle ultrasound models on external data

Published: 15 August 2025| Version 2 | DOI: 10.17632/7bcsvh6nfv.2
Contributors:
Michał Błaż, Monika Ostrowska, Agnieszka Kułaga, Ewa Maludzińska, Michał Michalski

Description

This repository contains model weights that were fine-tuned (using uploaded configs) from the original muscle segmentation models described in the paper "Machine learning-driven Heckmatt grading in facioscapulohumeral muscular dystrophy: A novel pathway for musculoskeletal ultrasound analysis" by Marzola et al. DOI: 10.1016/j.clinph.2025.01.016 The framework and idea behind this work can be found in the paper "Evaluating machine learning muscle ultrasound models on external data – is this the way to propagate forward?" by Błaż et al. DOI: 10.1016/j.clinph.2025.2110968 This repository does not contain any muscle ultrasound images.

Files

Steps to reproduce

Please follow instructions at: https://github.com/blazuu/muscle_ultrasound

Institutions

  • Krakowski Szpital Specjalistyczny im Jana Pawla II

Categories

Ultrasonics, Image Segmentation, Muscle Disorder, Ultrasound Imaging of the Muscles

Licence