arXiv:2605.27796v1 Announce Type: cross 
Abstract: Ultrasound is widely used in obstetric care due to its safety, accessibility, and real-time imaging. However, interpretation remains operator-dependent and susceptible to noise and artifacts. Deep learning models have shown strong performance to solve these problem, but they typically require large annotated datasets that are difficult to obtain in clinical ultrasound. Foundation models (FMs) offer an alternative, using a large number of ultrasound images to learn transferable representations that can generalize with limited labeled data. This work presents a comprehensive benchmark of ultrasound-specific FMs for fetal plane classification. We evaluated four ultrasound FMs (USFM, MOFO, UltraSAM, FetalCLIP) against two CNN baselines (ResNet50, EfficientNet-V2) and a ViT (DINOv3) pretrained on natural images. We trained all models under two complementary settings: full fine-tuning and linear probing with a frozen encoder. All models were trained using 5-fold patient-level cross-validation on a Spanish fetal ultrasound dataset and tested on both in-domain data and an external African cohort to assess cross-population generalization. We found that FetalCLIP achieved the best results in the linear probing setting (F1 = 0.9261 for in-domain, F1 = 0.9731 for out-of-domain), while USFM performed best in the full fine-tuning setting (F1 = 0.9476 for in-domain, F1 = 0.9515 for out-of-domain). MOFO and UltraSAM degraded most in both settings, underperforming natural image pretrained models in some cases. These findings highlight how the choice of pretrained model strongly affects fetal plane classification performance, since different pretraining objectives lead to different levels of transferability.

تم إجراء تقييم شامل لنماذج الأساس بالموجات فوق الصوتية (FMs) لتصنيف الطائرات الجنينية، حيث تم تقييم أربعة نماذج—USFM وMOFO وUltraSAM وFetalCLIP—مقابل نماذج CNN المعروفة مثل ResNet50 وEfficientNet-V2. تسلط الدراسة الضوء على إمكانية هذه النماذج في تحسين تفسير الصور الجنينية، والتي غالبًا ما تعيقها الاعتماد على المشغل والضوضاء.

Se ha realizado un benchmark integral sobre modelos de fundación de ultrasonido (FMs) para la clasificación de planos fetales, evaluando cuatro modelos—USFM, MOFO, UltraSAM y FetalCLIP—frente a bases CNN establecidas como ResNet50 y EfficientNet-V2. El estudio destaca el potencial de estos modelos para mejorar la interpretación de imágenes fetales, que a menudo se ve obstaculizada por la dependencia del operador y el ruido.

Un benchmark complet a été réalisé sur des modèles de fondation d'échographie (FMs) pour la classification des plans fœtaux, évaluant quatre modèles—USFM, MOFO, UltraSAM et FetalCLIP—contre des bases CNN établies comme ResNet50 et EfficientNet-V2. L'étude met en lumière le potentiel de ces modèles pour améliorer l'interprétation des images fœtales, souvent entravée par la dépendance à l'opérateur et le bruit.

A comprehensive benchmark has been conducted on ultrasound foundation models (FMs) for fetal plane classification, evaluating four models—USFM, MOFO, UltraSAM, and FetalCLIP—against established CNN baselines like ResNet50 and EfficientNet-V2. The study highlights the potential of these models to improve fetal imaging interpretation, which is often hindered by operator dependency and noise.