Comparative Analysis of 3D Convolutional and 2.5D Slice-Conditioned U-Net Architectures for MRI Super-Resolution via Elucidated Diffusion Models

TL;DR

This study compares 3D convolutional and 2.5D slice-conditioned U-Net architectures within an elucidated diffusion model framework for MRI super-resolution, demonstrating superior performance of the 3D model on brain MRI data.

Contribution

It introduces a comparative analysis of 3D and 2.5D U-Net architectures for MRI super-resolution using diffusion models, highlighting the effectiveness of 3D processing.

Findings

3D U-Net outperforms 2.5D in PSNR, SSIM, and LPIPS metrics

The 3D model surpasses pretrained EDSR baseline in all metrics

Diffusion-based MRI super-resolution benefits from volumetric processing

Abstract

Magnetic resonance imaging (MRI) super-resolution (SR) methods that computationally enhance low-resolution acquisitions to approximate high-resolution quality offer a compelling alternative to expensive high-field scanners. In this work we investigate an elucidated diffusion model (EDM) framework for brain MRI SR and compare two U-Net backbone architectures: (i) a full 3D convolutional U-Net that processes volumetric patches with 3D convolutions and multi-head self-attention, and (ii) a 2.5D slice-conditioned U-Net that super-resolves each slice independently while conditioning on an adjacent slice for inter-slice context. Both models employ continuous-sigma noise conditioning following Karras et al. and are trained on the NKI cohort of the FOMO60K dataset. On a held-out test set of 5 subjects (6 volumes, 993 slices), the 3D model achieves 37.75 dB PSNR, 0.997 SSIM, and 0.020 LPIPS, improving on the off-the-shelf pretrained EDSR baseline (35.57 dB / 0.024 LPIPS) and the 2.5D variant (35.82 dB) across all three metrics under the same test data and degradation pipeline.