FetalFlex: Anatomy-Guided Diffusion Model for Flexible Control on Fetal Ultrasound Image Synthesis

2025-03-19

AI Summary

Fetal ultrasound multiplanar annotation data—particularly for rare anomalies—are severely scarce, hindering AI model training and clinical education. To address this, we propose an anatomy-guided controllable diffusion generation framework that enables out-of-distribution fetal ultrasound image synthesis without requiring real abnormal samples. Our method integrates anatomical prior encoding, a pre-alignment module, Repaint-based texture-consistent inpainting, and a two-stage adaptive sampling strategy to ensure cross-plane anatomical consistency and pathologically plausible abnormalities. Evaluated on multicenter datasets, our approach achieves state-of-the-art image quality (FID reduced by 23.6%, LPIPS reduced by 18.4%) and significantly improves downstream anomaly detection across six models (average AUC increased by 5.2%). Clinical credibility is further validated through blinded expert radiologist assessment.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Fetal ultrasound (US) examinations require the acquisition of multiple planes, each providing unique diagnostic information to evaluate fetal development and screening for congenital anomalies. However, obtaining a comprehensive, multi-plane annotated fetal US dataset remains challenging, particularly for rare or complex anomalies owing to their low incidence and numerous subtypes. This poses difficulties in training novice radiologists and developing robust AI models, especially for detecting abnormal fetuses. In this study, we introduce a Flexible Fetal US image generation framework (FetalFlex) to address these challenges, which leverages anatomical structures and multimodal information to enable controllable synthesis of fetal US images across diverse planes. Specifically, FetalFlex incorporates a pre-alignment module to enhance controllability and introduces a repaint strategy to ensure consistent texture and appearance. Moreover, a two-stage adaptive sampling strategy is developed to progressively refine image quality from coarse to fine levels. We believe that FetalFlex is the first method capable of generating both in-distribution normal and out-of-distribution abnormal fetal US images, without requiring any abnormal data. Experiments on multi-center datasets demonstrate that FetalFlex achieved state-of-the-art performance across multiple image quality metrics. A reader study further confirms the close alignment of the generated results with expert visual assessments. Furthermore, synthetic images by FetalFlex significantly improve the performance of six typical deep models in downstream classification and anomaly detection tasks. Lastly, FetalFlex's anatomy-level controllable generation offers a unique advantage for anomaly simulation and creating paired or counterfactual data at the pixel level. The demo is available at: https://dyf1023.github.io/FetalFlex/.

Problem

• Generates diverse fetal ultrasound images for training and AI development.
• Enhances image synthesis control using anatomical structures and multimodal data.
• Improves anomaly detection and classification in fetal ultrasound imaging.

Innovation

• Anatomy-guided diffusion model for fetal US synthesis
• Pre-alignment module enhances image controllability
• Two-stage adaptive sampling refines image quality