Diffusing the Blind Spot: Uterine MRI Synthesis with Diffusion Models

2025-08-11

AI Summary

Existing diffusion models struggle to generate anatomically accurate female pelvic MRI images, limiting their utility in gynecological imaging—where data scarcity and privacy sensitivity are critical concerns. To address this, we propose the first latent-space diffusion framework specifically designed for uterine MRI synthesis, integrating Latent Diffusion Models (LDM) with Denoising Diffusion Probabilistic Models (DDPM) to support unconditional and conditional 2D/3D generation. We enhance anatomical fidelity via perceptual loss and distribution-matching regularization, and validate clinical realism through blinded radiologist assessment. We release the first privacy-preserving synthetic uterine MRI dataset—including multiplanar T2-weighted images—enabling reproducible research. Our synthetic data significantly improves downstream classification accuracy (+8.2%) over conventional reconstruction methods. Both code and dataset are publicly available, advancing equitable, transparent, and clinically grounded AI research in gynecology.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Privacy
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Security and Privacy: Data transparency and provenance
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Despite significant progress in generative modelling, existing diffusion models often struggle to produce anatomically precise female pelvic images, limiting their application in gynaecological imaging, where data scarcity and patient privacy concerns are critical. To overcome these barriers, we introduce a novel diffusion-based framework for uterine MRI synthesis, integrating both unconditional and conditioned Denoising Diffusion Probabilistic Models (DDPMs) and Latent Diffusion Models (LDMs) in 2D and 3D. Our approach generates anatomically coherent, high fidelity synthetic images that closely mimic real scans and provide valuable resources for training robust diagnostic models. We evaluate generative quality using advanced perceptual and distributional metrics, benchmarking against standard reconstruction methods, and demonstrate substantial gains in diagnostic accuracy on a key classification task. A blinded expert evaluation further validates the clinical realism of our synthetic images. We release our models with privacy safeguards and a comprehensive synthetic uterine MRI dataset to support reproducible research and advance equitable AI in gynaecology.

Problem

• Generating anatomically precise female pelvic MRI images
• Overcoming data scarcity and privacy in gynaecological imaging
• Improving diagnostic accuracy with synthetic uterine MRI data

Innovation

• Diffusion-based framework for uterine MRI synthesis
• Combines unconditional and conditioned DDPMs and LDMs
• Generates high fidelity, anatomically coherent synthetic images