Reconstruct or Generate: Exploring the Spectrum of Generative Modeling for Cardiac MRI

2025-07-25

AI Summary

This work investigates the performance trade-offs between reconstruction (e.g., image inpainting, super-resolution) and generation (e.g., synthetic data augmentation, counterfactual analysis) tasks for generative models in cardiac MRI. We propose the “reconstruction–generation continuum” framework and systematically evaluate diffusion and autoregressive models under varying mask ratios and sampling strategies, establishing the first medical imaging–specific “Generative Model Zoo” benchmark. Experiments reveal that diffusion models achieve superior perceptual quality in unconditional generation but suffer from hallucinations under high masking rates; in contrast, autoregressive models demonstrate greater robustness and stability in reconstruction tasks. Our analysis uncovers a fundamental fidelity–perception trade-off distinguishing these model families, offering new theoretical insights and empirically grounded guidelines for task-aware selection of generative models in medical imaging.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

In medical imaging, generative models are increasingly relied upon for two distinct but equally critical tasks: reconstruction, where the goal is to restore medical imaging (usually inverse problems like inpainting or superresolution), and generation, where synthetic data is created to augment datasets or carry out counterfactual analysis. Despite shared architecture and learning frameworks, they prioritize different goals: generation seeks high perceptual quality and diversity, while reconstruction focuses on data fidelity and faithfulness. In this work, we introduce a "generative model zoo" and systematically analyze how modern latent diffusion models and autoregressive models navigate the reconstruction-generation spectrum. We benchmark a suite of generative models across representative cardiac medical imaging tasks, focusing on image inpainting with varying masking ratios and sampling strategies, as well as unconditional image generation. Our findings show that diffusion models offer superior perceptual quality for unconditional generation but tend to hallucinate as masking ratios increase, whereas autoregressive models maintain stable perceptual performance across masking levels, albeit with generally lower fidelity.

Problem

• Exploring generative models for cardiac MRI reconstruction and generation
• Comparing diffusion and autoregressive models in medical imaging tasks
• Analyzing trade-offs between perceptual quality and data fidelity

Innovation

• Generative model zoo for cardiac MRI analysis
• Latent diffusion models for superior perceptual quality
• Autoregressive models for stable perceptual performance