MRI Super-Resolution with RCDM/WaveMix and Task-Aware Segmentation

2026-09-30

AI Summary

This study addresses the unclear impact of super-resolution enhancement of 1.5T brain MRI on downstream tissue segmentation. We propose a lightweight, physics-guided unpaired unsupervised training framework that incorporates an identity-penalized cycle consistency objective to enhance images by adding scanner-specific textures without altering anatomical structures. By integrating residual adversarial networks, stochastic degradation operators, and wavelet token mixing, we systematically evaluate its effects on segmentation models with distinct architectures, including U-Net and Swin-UNet. Our findings reveal a model-dependent nature of the enhancement: it significantly improves the Dice coefficient for the wavelet-based segmenter yet degrades U-Net performance. This underscores the necessity of validating enhancement efficacy for specific models using reliable annotations.

Abstract

Super-resolution and quality enhancement of 1.5\,T brain MRI are normally validated with image-fidelity metrics, although their purpose is to improve downstream analysis. We study whether enhancement improves tissue segmentation, and for which segmenters. We propose an unpaired, physics-guided training pipeline for a lightweight ($\le$2.5\,M parameter) recurrent convolutional enhancer: a six-module stochastic 1.5\,T degradation operator, a residual adversarial network that adds scanner-specific texture without moving anatomy, and a cycle-consistent objective with an anti-identity penalty that rules out the copy solution. We then train U-Net, Swin-UNet and wavelet token-mixing segmenters \citep{jeevan2023wavemix} from scratch on either raw or enhanced 1.5\,T images of the same subjects, using identical labels and subject-level splits, for three enhancer variants and two datasets. On ABIDE (41 held-out subjects, FreeSurfer labels) enhancement significantly improves the wavelet segmenter (mean Dice $+0.014$, Wilcoxon $p=3.5\times10^{-5}$; CSF $+0.018$, grey matter $+0.013$), significantly degrades the U-Net ($-0.008$, $p=5.1\times10^{-4}$) and leaves Swin-UNet unchanged. On IXI, whose labels come from FSL-FAST, enhancement lowers Dice for all nine pairings, almost entirely through CSF; we trace this to spatially implausible CSF voxels in the labels that penalise smoother predictions. Enhancement of low-field MRI should therefore be validated per downstream model and against reliable labels.

Problem

• MRI Super-Resolution
• Downstream Segmentation
• Image Enhancement Validation
• Task-Aware Evaluation

Innovation

• Super-resolution
• Physics-guided training
• Cycle-consistent objective
• Task-aware segmentation
• Wavelet token-mixing