CD-DPE: Dual-Prompt Expert Network Based on Convolutional Dictionary Feature Decoupling for Multi-Contrast MRI Super-Resolution

TL;DR

This work tackles multi-contrast MRI super-resolution by introducing CD-DPE, a dual-prompt expert network that decouples features into cross-contrast unique and intra-contrast common components using a convolutional dictionary framework (CD-FDM) and then adaptively fuses them via a frequency-aware, routing-based module (DP-FFEM). The method achieves state-of-the-art SR performance on BraTS2018 and IXI datasets for 2x and 4x upsampling, and demonstrates strong generalization to unseen data. Ablation studies confirm the critical roles of CD-FDM and DP-FFEM, as well as the importance of the consistency and decoupling losses. The results indicate CD-DPE's potential for reliable transfer of structural information across contrasts, with practical implications for faster MRI protocols and improved diagnostic detail.

Abstract

Multi-contrast magnetic resonance imaging (MRI) super-resolution intends to reconstruct high-resolution (HR) images from low-resolution (LR) scans by leveraging structural information present in HR reference images acquired with different contrasts. This technique enhances anatomical detail and soft tissue differentiation, which is vital for early diagnosis and clinical decision-making. However, inherent contrasts disparities between modalities pose fundamental challenges in effectively utilizing reference image textures to guide target image reconstruction, often resulting in suboptimal feature integration. To address this issue, we propose a dual-prompt expert network based on a convolutional dictionary feature decoupling (CD-DPE) strategy for multi-contrast MRI super-resolution. Specifically, we introduce an iterative convolutional dictionary feature decoupling module (CD-FDM) to separate features into cross-contrast and intra-contrast components, thereby reducing redundancy and interference. To fully integrate these features, a novel dual-prompt feature fusion expert module (DP-FFEM) is proposed. This module uses a frequency prompt to guide the selection of relevant reference features for incorporation into the target image, while an adaptive routing prompt determines the optimal method for fusing reference and target features to enhance reconstruction quality. Extensive experiments on public multi-contrast MRI datasets demonstrate that CD-DPE outperforms state-of-the-art methods in reconstructing fine details. Additionally, experiments on unseen datasets demonstrated that CD-DPE exhibits strong generalization capabilities.

Figures (9)

• Figure 1: Structural disparities and shared information across multi-contrast MRI.
• Figure 2: The architecture of dual-prompt expert network based on convolutional dictionary feature decoupling (CD-DPE).
• Figure 3: The architecture of dual-prompt feature fusion expert module (DP-FFEM).
• Figure 4: Qualitative comparison of various methods on the BraTS2018 dataset with $4\times$ SR. The yellow arrows indicate areas with significant differences, which are enlarged and shown with residual plots compared to ground-truth (GT).
• Figure 5: Qualitative comparison of various methods on the IXI dataset with $4\times$ SR. The yellow arrows indicate areas with significant differences, which are enlarged and shown with residual plots compared to GT.