MRI Image Generation Based on Text Prompts

2025-05-23

AI Summary

Medical MRI data acquisition is hindered by high costs, scarcity of rare-disease samples, and strict privacy constraints. To address these challenges, this work proposes a text-guided cross-field-strength MRI synthesis method. Leveraging a fine-tuned Stable Diffusion framework and jointly trained on fastMRI and M4Raw datasets, it achieves the first text-to-image generation of multi-contrast (T1, T2, FLAIR) brain MRI, with controllable synthesis across 0.3T and 3T field strengths. Quantitative evaluation using FID and MS-SSIM confirms high fidelity and semantic consistency of synthesized images. Furthermore, few-shot classification experiments demonstrate substantial downstream utility: classification accuracy improves by 12.6% on a small 0.35T dataset when augmented with synthetic data. This work establishes a novel paradigm for privacy-preserving, cost-efficient, and generalizable medical AI data generation.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

This study explores the use of text-prompted MRI image generation with the Stable Diffusion (SD) model to address challenges in acquiring real MRI datasets, such as high costs, limited rare case samples, and privacy concerns. The SD model, pre-trained on natural images, was fine-tuned using the 3T fastMRI dataset and the 0.3T M4Raw dataset, with the goal of generating brain T1, T2, and FLAIR images across different magnetic field strengths. The performance of the fine-tuned model was evaluated using quantitative metrics,including Fr'echet Inception Distance (FID) and Multi-Scale Structural Similarity (MS-SSIM), showing improvements in image quality and semantic consistency with the text prompts. To further evaluate the model's potential, a simple classification task was carried out using a small 0.35T MRI dataset, demonstrating that the synthetic images generated by the fine-tuned SD model can effectively augment training datasets and improve the performance of MRI constrast classification tasks. Overall, our findings suggest that text-prompted MRI image generation is feasible and can serve as a useful tool for medical AI applications.

Problem

• Generate MRI images from text prompts to overcome data scarcity
• Fine-tune Stable Diffusion model for diverse MRI contrast types
• Enhance medical AI training with synthetic MRI dataset augmentation

Innovation

• Fine-tuned Stable Diffusion for MRI generation
• Text-prompted brain image synthesis
• Synthetic data improves classification performance