FetalCLIP: A Visual-Language Foundation Model for Fetal Ultrasound Image Analysis

TL;DR

FetalCLIP is a large vision-language foundation model trained on over 210,000 fetal ultrasound images and text, enabling robust, generalizable analysis for various fetal ultrasound tasks.

Contribution

This work introduces FetalCLIP, the largest paired dataset and a novel multimodal foundation model specifically designed for fetal ultrasound image analysis.

Findings

Outperforms baseline models across multiple fetal ultrasound tasks

Demonstrates strong generalization with limited labeled data

Achieves state-of-the-art results in classification, gestational age estimation, CHD detection, and segmentation

Abstract

Foundation models are becoming increasingly effective in the medical domain, offering pre-trained models on large datasets that can be readily adapted for downstream tasks. Despite progress, fetal ultrasound images remain a challenging domain for foundation models due to their inherent complexity, often requiring substantial additional training and facing limitations due to the scarcity of paired multimodal data. To overcome these challenges, here we introduce FetalCLIP, a vision-language foundation model capable of generating universal representation of fetal ultrasound images. FetalCLIP was pre-trained using a multimodal learning approach on a diverse dataset of 210,035 fetal ultrasound images paired with text. This represents the largest paired dataset of its kind used for foundation model development to date. This unique training approach allows FetalCLIP to effectively learn the intricate anatomical features present in fetal ultrasound images, resulting in robust representations that can be used for a variety of downstream applications. In extensive benchmarking across a range of key fetal ultrasound applications, including classification, gestational age estimation, congenital heart defect (CHD) detection, and fetal structure segmentation, FetalCLIP outperformed all baselines while demonstrating remarkable generalizability and strong performance even with limited labeled data. We plan to release the FetalCLIP model publicly for the benefit of the broader scientific community.