DGCM-SSFA: Dual-Granularity Cross-Modulation and Sparse Strip Frequency Attention for Efficient 3T-to-7T MRI Synthesis

Zhichun Fan; Zhiqiang Chen; Lanling Liu; Lingjun Liu; Zhonghua Xie

cs.CVOct 21, 2026

DGCM-SSFA: Dual-Granularity Cross-Modulation and Sparse Strip Frequency Attention for Efficient 3T-to-7T MRI Synthesis

Authors: Zhichun Fan, Zhiqiang Chen, Lanling Liu, Lingjun Liu, Zhonghua Xie

Organizations: School of Computer Science and Engineering, Huizhou University, Huizhou, China

Abstract

Synthesizing high-fidelity 7T MRI from routinely acquired 3T scans offers a practical pathway to enhanced neuroimaging without the prohibitive cost of ultra-high-field scanners. While existing deep learning approaches have achieved promising results, state-of-the-art methods typically rely on either large-scale architectures exceeding 45 million parameters or computationally intensive ensemble strategies, limiting their practicality in resource-constrained clinical environments. In this work, we propose a compact multi-domain attention framework, termed DGCM-SSFA, that achieves competitive synthesis quality with only 11M parameters. The proposed architecture integrates two complementary modules into a U-Net Transformer backbone. First, a Dual-Granularity Cross-Modulated Feature Modulator (DGCM) is embedded at multiple hierarchical levels to enhance local edge and texture details through Sobel-guided spatial attention and depthwise frequency-aware channel attention with cross-branch gating. Second, a Sparse Strip Frequency Attention (SSFA) module at the bottleneck captures global spatial context via strip-pooled attention with near-linear complexity, while simultaneously performing FFT-based spectral modulation in the frequency domain. These two modules operate synergistically with the backbone's native channel attention, forming a triple-granularity feature refinement pipeline. Extensive experiments on the UNC and BNU 3T-to-7T MRI datasets across T1w and T2w modalities demonstrate that DGCM-SSFA achieves strong structural fidelity, attaining an SSIM of 0.8431 on BNU T2w---approaching the best benchmark value of 0.8624---and 0.8100 on BNU T1w, while using substantially fewer parameters than Transformer-based alternatives such as ResViT (123.4M) and FS-RWKV (45.4M).

Figures & tables

Fig. 1: Overall architecture of the proposed DGCM-SSFA. (a) The framework adopts a U-Net-style encoder-decoder backbone. The encoder stacks FFN and Conv layers for hierarchical feature extraction. DGCM modules are inserted after each encoder level and after the refinement stage for local detail enhancement. A single SSFA module is deployed at the bottleneck for global spatial-frequency attention. A global residual connection produces the final output. (b) DGCM consists of an edge-aware spatial branch and a frequency-aware channel branch with cross-modulation fusion. (c) SSFA combines sparse strip attention (SSA) and frequency spectral attention (FSA) with cross-gated fusion.

Method	UNC T1w			UNC T2w			BNU T1w			BNU T2w			Params (M)
	PSNR $\uparrow$	SSIM $\uparrow$	RMSE $\downarrow$	PSNR $\uparrow$	SSIM $\uparrow$	RMSE $\downarrow$	PSNR $\uparrow$	SSIM $\uparrow$	RMSE $\downarrow$	PSNR $\uparrow$	SSIM $\uparrow$	RMSE $\downarrow$
CycleGAN	19.98	0.6789	0.1008	24.88	0.7538	0.0586	21.93	0.7881	0.0807	26.30	0.8053	0.0493	28.29
pix2pix	19.96	0.6847	0.1015	24.74	0.7550	0.0591	22.74	0.8102	0.0739	26.47	0.8182	0.0481	57.18
DDCNN	20.75	0.6412	0.0924	24.21	0.7286	0.0640	23.02	0.7732	0.0717	26.88	0.8229	0.0463	0.29
WATNet	20.17	0.6753	0.0990	23.07	0.6698	0.0719	21.13	0.7319	0.0886	26.13	0.7625	0.0503	1.35
RegGAN	19.46	0.6423	0.1073	23.97	0.6994	0.0654	22.10	0.7897	0.0793	25.61	0.7703	0.0531	2.76

TABLE I: Quantitative Comparison on MRI Cross-Modality Synthesis (3T

\rightarrow

7T)

Fig. 2: Qualitative comparison of synthesized 3T-to-7T MRI on four datasets (UNC T1w, UNC T2w, BNU T1w, BNU T2w). From left to right: 3T input, 7T ground truth, CycleGAN, DDCNN, pix2pix, FS-RWKV, the proposed DGCM-SSFA, and error map (error magnitude relative to GT). Yellow and cyan boxes in the main image denote regions magnified below.

Variant	PSNR $\uparrow$	SSIM $\uparrow$
Full DGCM-SSFA	26.27	0.8431
w/o DGCM	25.88	0.8206
w/o SSFA	26.02	0.8270
w/o Both	25.66	0.8048

TABLE II: Ablation study on the key components of DGCM-SSFA. All models are trained without ensemble and evaluated on BNU T2w. The best result in each column is highlighted in bold.

Fig. 3: Ablation study on the key components of DGCM-SSFA. PSNR (left) and SSIM (right) are reported for four variants. All models are trained without ensemble and evaluated on BNU T2w.

Medical Image Synthesis

Medical image synthesis creates, translates, reconstructs, and edits clinical images to fill missing data and support analysis.

See all 312 papers

Jul 18, 2025

Converting T1-weighted MRI from 3T to 7T quality using deep learning

Ultra-high resolution 7 tesla (7T) magnetic resonance imaging (MRI) provides detailed anatomical views, offering better signal-to-noise ratio, resolution and tissue contrast than 3T MRI, though at the cost of accessibility. We present an advanced deep learning model for synthesizing 7T brain MRI from 3T brain MRI. Paired 7T and 3T T1-weighted images were acquired from 172 participants (124 cognitively unimpaired, 48 impaired) from the Swedish BioFINDER-2 study. To synthesize 7T MRI from 3T images, we trained two models: a specialized U-Net, and a U-Net integrated with a generative adversarial network (GAN U-Net). Our models outperformed two previous state-of-the-art 3T-to-7T models in image-based evaluation metrics. Four blinded MRI professionals judged our synthetic 7T images as comparable in detail to real 7T images, and superior in subjective visual quality to 7T images, due to the reduction of artifacts. Using both SynthSeg and NextBrain, automated segmentations of the synthetic 7T images were more similar to real 7T segmentations than automated segmentations from the 3T images that were used to synthesize the 7T images. Finally, synthetic 7T images showed similar performance to real 3T images in downstream prediction of cognitive status using MRI derivatives (n=3,168). In all, we show that synthetic T1-weighted brain images approaching 7T quality can be generated from 3T images, which may improve image quality and segmentation, without compromising performance in downstream tasks. Future directions, possible clinical use cases, and limitations are discussed.

Malo Gicquel, Ruoyi Zhao, Anika Wuestefeld +12

Department of Clinical Sciences Malmö, SciLifeLab, Lund University, Lund, Sweden · Univ Rennes, CNRS, Inria, Inserm, IRISA UMR 6074, Empenn ERL U 1228, Rennes, France · Clinical Memory Research Unit, Department of Clinical Sciences Malmö, Lund University, Lund, Sweden +8

Magnetic Resonance Imaging Medical Image Synthesis Brain MRI

Ask PDF

Apr 22, 2026

WFM: 3D Wavelet Flow Matching for Ultrafast Multi-Modal MRI Synthesis

Diffusion models have achieved remarkable quality in multi-modal MRI synthesis, but their computational cost (hundreds of sampling steps and separate models per modality) limits clinical deployment. We observe that this inefficiency stems from an unnecessary starting point: diffusion begins from pure noise, discarding the structural information already present in available MRI sequences. We propose WFM (Wavelet Flow Matching), which instead learns a direct flow from an informed prior, the mean of conditioning modalities in wavelet space, to the target distribution. Because the source and target share underlying anatomy and differ primarily in contrast, this formulation enables accurate synthesis in just 1-2 integration steps. A single 82M-parameter model with class conditioning synthesizes all four BraTS modalities (T1, T1c, T2, FLAIR), replacing four separate diffusion models totaling 326M parameters. On BraTS 2024, WFM achieves 26.8 dB PSNR and 0.94 SSIM, within 1-2 dB of diffusion baselines, while running 250-1000x faster (0.16-0.64s vs. 160s per volume). This speed-quality trade-off makes real-time MRI synthesis practical for clinical workflows. Code is available at https://github.com/yalcintur/WFM.

Yalcin Tur, Mihajlo Stojkovic, Ulas Bagci

Department of Computer Science, Stanford University, Stanford, CA, USA · Machine and Hybrid Intelligence Lab, Feinberg School of Medicine, Northwestern University, Chicago, IL, USA

Flow Matching BraTS Magnetic Resonance Imaging Medical Image Synthesis

Ask PDF

Apr 1, 2026

Automated Detection of Multiple Sclerosis Lesions on 7-tesla MRI Using U-net and Transformer-based Segmentation

Ultra-high field 7-tesla (7T) MRI improves visualization of multiple sclerosis (MS) white matter lesions (WML) but differs sufficiently in contrast and artifacts from 1.5-3T imaging - suggesting that widely used automated segmentation tools may not translate directly. We analyzed 7T FLAIR scans and generated reference WML masks from Lesion Segmentation Tool (LST) outputs followed by expert manual revision. As external comparators, we applied LST-LPA and the more recent LST-AI ensemble, both originally developed on lower-field data. We then trained 3D UNETR and SegFormer transformer-based models on 7T FLAIR at multiple resolutions (0.5x0.5x0.5^3, 1.0x1.0x1.0^3, and 1.5x1.5x2.0^3) and evaluated all methods using voxel-wise and lesion-wise metrics from the BraTS 2023 framework. On the held-out test set at native 0.5x0.5x0.5^3 resolution, 7T-trained transformers achieved competitive overlap with LST-AI while recovering additional small lesions that were missed by classical methods, at the cost of some boundary variability and occasional artifact-related false positives. On a held-out 7 T test set, our best transformer model (SegFormer) achieved a voxel-wise Dice of 0.61 and lesion-wise Dice of 0.20, improving on the classical LST-LPA tool (Dice 0.39, lesion-wise Dice 0.02). Performance decreased for models trained on downsampled images, underscoring the value of native 7T resolution for small-lesion detection. By releasing our 7T-trained models, we aim to provide a reproducible, ready-to-use resource for automated lesion quantification in ultra-high field MS research (https://github.com/maynord/7T-MS-lesion-segmentation).

Michael Maynord, Minghui Liu, Cornelia Fermüller +4

Department of Computer, Science University of Maryland College Park, College Park, 20742, MD, USA · Department of Neurology, University of Maryland School of Medicine, Baltimore, MD, 21201, USA · Baltimore VA Medical Center, VA Maryland Healthcare System, Baltimore, MD, 21201, USA

Medical Image Segmentation 3D Medical Image Segmentation Brain MRI Semantic Segmentation Magnetic Resonance Imaging

Ask PDF

Abstract

Medical Image Synthesis

Explore similar work

Converting T1-weighted MRI from 3T to 7T quality using deep learning

WFM: 3D Wavelet Flow Matching for Ultrafast Multi-Modal MRI Synthesis

Automated Detection of Multiple Sclerosis Lesions on 7-tesla MRI Using U-net and Transformer-based Segmentation