Benchmarking GANs, Diffusion Models, and Flow Matching for T1w-to-T2w MRI Translation

2025-07-19

AI Summary

This study addresses the clinical challenge of prolonged acquisition time and high cost associated with multimodal MRI scanning by systematically evaluating generative adversarial networks (GANs), diffusion models, and flow-matching methods for T1-weighted-to-T2-weighted (T1w→T2w) cross-modal medical image synthesis. For the first time, we conduct a fair and comprehensive comparison of these three dominant generative paradigms under unified experimental settings across three public MRI datasets. Results demonstrate that Pix2Pix GAN achieves superior structural fidelity (higher SSIM), image quality (higher PSNR), and inference efficiency (3–5× faster) compared to diffusion and flow-matching approaches—particularly exhibiting greater robustness in low-data regimes. In contrast, flow-based models show pronounced overfitting tendencies. To foster reproducibility and clinical deployment, we release all code and pretrained models. This work establishes an empirical benchmark and practical guidance for lightweight, reliable, and deployable medical image synthesis.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Multimodal Learning
• Natural Language Processing: Generation

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Magnetic Resonance Imaging (MRI) enables the acquisition of multiple image contrasts, such as T1-weighted (T1w) and T2-weighted (T2w) scans, each offering distinct diagnostic insights. However, acquiring all desired modalities increases scan time and cost, motivating research into computational methods for cross-modal synthesis. To address this, recent approaches aim to synthesize missing MRI contrasts from those already acquired, reducing acquisition time while preserving diagnostic quality. Image-to-image (I2I) translation provides a promising framework for this task. In this paper, we present a comprehensive benchmark of generative models$unicode{x2013}$specifically, Generative Adversarial Networks (GANs), diffusion models, and flow matching (FM) techniques$unicode{x2013}$for T1w-to-T2w 2D MRI I2I translation. All frameworks are implemented with comparable settings and evaluated on three publicly available MRI datasets of healthy adults. Our quantitative and qualitative analyses show that the GAN-based Pix2Pix model outperforms diffusion and FM-based methods in terms of structural fidelity, image quality, and computational efficiency. Consistent with existing literature, these results suggest that flow-based models are prone to overfitting on small datasets and simpler tasks, and may require more data to match or surpass GAN performance. These findings offer practical guidance for deploying I2I translation techniques in real-world MRI workflows and highlight promising directions for future research in cross-modal medical image synthesis. Code and models are publicly available at https://github.com/AndreaMoschetto/medical-I2I-benchmark.

Problem

• Benchmarking GANs, diffusion models, and flow matching for MRI contrast synthesis
• Evaluating T1w-to-T2w translation performance in computational efficiency and image quality
• Addressing overfitting in flow-based models for small medical datasets

Innovation

• Benchmarking GANs, diffusion models, flow matching
• Pix2Pix GAN outperforms in MRI translation
• Publicly available code and models provided