Synergistic Foundation Models for Semi-Supervised Fetal Cardiac Ultrasound Analysis: SAM-Med2D Boundary Refinement and DINOv3 Semantic Enhancement

TL;DR

This paper introduces a semi-supervised framework combining boundary refinement and semantic enhancement for fetal cardiac ultrasound analysis, achieving high accuracy in segmentation and classification tasks.

Contribution

It proposes a novel joint segmentation and classification method using SAM-Med2D and DINOv3 with a two-stage optimization strategy for improved fetal ultrasound analysis.

Findings

Achieved Dice Similarity Coefficient of 79.99%

Normalized Surface Distance of 61.62%

F1-score of 41.20% on FETUS 2026 leaderboard

Abstract

We present a semi-supervised framework for joint segmentation and classification of fetal cardiac ultrasound images. Built upon the EchoCare multi-task backbone, our method integrates SAM-Med2D for boundary refinement and leverages DINOv3 to enhance pseudo-label quality. We introduce view-specific hard masking along with a two-stage optimization strategy: an EMA phase to consolidate segmentation capabilities, followed by a Classification Fine-Tuning phase that freezes segmentation parameters and resets the classification head to recover classification performance without compromising segmentation gains. Evaluated on the FETUS 2026 leaderboard, our method achieves a Dice Similarity Coefficient at 79.99%, Normalized Surface Distance at 61.62%, and F1-score at 41.20%, validating the effectiveness of our approach for prenatal congenital heart disease screening. Source code is publicly available at: https://github.com/2826056177/zcst_fetus2026.