Dual Agreement Consistency Learning with Foundation Models for Semi-Supervised Fetal Heart Ultrasound Segmentation and Diagnosis

TL;DR

This paper introduces FM-DACL, a semi-supervised learning framework that combines foundation models and co-training to improve fetal heart ultrasound segmentation and diagnosis with limited annotations.

Contribution

It proposes a novel semi-supervised dual agreement consistency learning method leveraging heterogeneous models for fetal cardiac ultrasound analysis.

Findings

Achieved a Dice score of 59.66 on the challenge dataset.

Demonstrated the effectiveness of heterogeneous co-training.

Validated the framework's potential for low-annotation scenarios.

Abstract

Congenital heart disease (CHD) screening from fetal echocardiography requires accurate analysis of multiple standard cardiac views, yet developing reliable artificial intelligence models remains challenging due to limited annotations and variable image quality. In this work, we propose FM-DACL, a semi-supervised Dual Agreement Consistency Learning framework for the FETUS 2026 challenge on fetal heart ultrasound segmentation and diagnosis. The method combines a pretrained ultrasound foundation model (EchoCare) with a convolutional network through heterogeneous co-training and an exponential moving average teacher to better exploit unlabeled data. Experiments on the multi-center challenge dataset show that FM-DACL achieves a Dice score of 59.66 and NSD of 42.82 using heterogeneous backbones, demonstrating the feasibility of the proposed semi-supervised framework. These results suggest that FM-DACL provides a flexible approach for leveraging heterogeneous models in low-annotation fetal cardiac ultrasound analysis. The code is available on https://github.com/13204942/FM-DACL.