South-Central Minzu University,a. College of Electronic and Information Engineering; b. Hubei Key Lab of Intelligent Wireless Communication; c. College of Computer Science,Wuhan 430074,China
Transformer-based MRI super-resolution methods offer strong global modeling capabilities but often overlook the role of deep prior constraints. To address this problem, a brain MRI super-resolution method based on diffusion priors is proposed, where a latent diffusion model generates structural priors to guide the Transformer in restoring fine details. A two-stage training strategy is adopted: the first stage constructs a content prior from ground-truth latent encodings to pretrain the reconstruction network; the second stage introduces diffusion-based priors and jointly optimizes the denoising and reconstruction processes under unsupervised conditions. Additionally, depthwise separable convolutions and permuted self-attention are employed to enhance modeling efficiency and expand the receptive field. Experiments on the IXI multi-modal MRI dataset (4×SR) demonstrate superior reconstruction quality and efficiency of the method over existing methods.
LIY, SIXOUB, PEYRINF. A review of the deep learning methods for medical images super resolution problems[J]. IRBM, 2021, 42(2): 120-133.
[2]
WANGZ, CHENJ, HOIS C H. Deep learning for image super-resolution: A survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(10): 3365-3387.
[3]
JIZ, ZOUB, KUIX, et al. Deep learning-based magnetic resonance image super-resolution: A survey[J]. Neural Computing and Applications, 2024, 36(21): 12725-12752.
[4]
DONGC, LOYC C, HEK, et al. Learning a deep convolutional network for image super-resolution[C]//Computer Vision-ECCV 2014. Cham: Springer, 2014: 184-199.
[5]
QIUD, ZHANGS, LIUY, et al. Super-resolution reconstruction of knee magnetic resonance imaging based on deep learning[J]. Computer Methods and Programs in Biomedicine, 2020, 187: 105059.
[6]
VASWANIA, SHAZEERN, PARMARN, et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30: 5998-6008.
[7]
DOSOVITSKIYA, BEYERL, KOLESNIKOVA, et al. An image is worth 16 × 16 words:Transformers for image recognition at scale[J]. arXiv: 2020,2010.11929.
[8]
LIUZ, LINY, CAOY, et al. Swin transformer: Hierarchical vision transformer using shifted windows[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Montreal:IEEE, 2022: 9992-10002.
[9]
YANC, SHIG, WUZ. SMIR: A Transformer-Based Model for MRI super-resolution reconstruction[C]//2021 IEEE International Conference on Medical Imaging Physics and Engineering (ICMIPE). Hefei:IEEE, 2022: 1-6.
[10]
FENGC M, YANY, FUH, et al. Task transformer network for joint MRI reconstruction and super-resolution[C]//Medical Image Computing and Computer Assisted Intervention-MICCAI 2021. Cham: Springer, 2021: 307-317.
[11]
FORIGUAC, ESCOBARM, ARBELAEZP. SuperFormer: Volumetric transformer architectures for MRI super-resolution[C]//Simulation and Synthesis in Medical Imaging. Cham: Springer, 2022: 132-141.
[12]
HUANGS, LIUX, TANT, et al. TransMRSR: Transformer-based self-distilled generative prior for brain MRI super-resolution[J]. The Visual Computer, 2023, 39(8): 3647-3659.
[13]
LIANGJ, CAOJ, SUNG, et al. SwinIR: Image restoration using swin transformer[C]//2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW).Montreal:IEEE, 2021: 1833-1844.
[14]
HO J, JAINA, ABBEELP. Denoising diffusion probabilistic models[C]//Proceedings of the 34th International Conference on Neural Information Processing Systems. Vancouver:ACM, 2020: 6840-6851.
[15]
CHUNGH, LEEE S, YEJ C. MR image denoising and super-resolution using regularized reverse diffusion[J]. IEEE Transactions on Medical Imaging, 2023, 42(4): 922-934.
[16]
WUZ, CHENX, XIES, et al. Super-resolution of brain MRI images based on denoising diffusion probabilistic model[J]. Biomedical Signal Processing and Control, 2023, 85: 104901.
[17]
ROMBACHR, BLATTMANNA, LORENZD, et al. High-resolution image synthesis with latent diffusion models[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New Orleans: IEEE, 2022: 10674-10685.
[18]
XIAB, ZHANGY, WANGS, et al. DiffIR: Efficient diffusion model for image restoration[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris:IEEE, 2024: 13049-13059.
[19]
CHENZ, ZHANGY, LIUD,et al.Hierarchical integration diffusion model for realistic image deblurring[J]// arXiv: 2023,2305.12966.
[20]
ZHOUY, LIZ, GUOC L, et al. SRFormer: Permuted self-attention for single image super-resolution[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV).Paris:IEEE, 2024: 12734-12745.
[21]
LIM B, SON S, KIMH, et al. Enhanced deep residual networks for single image super-resolution[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).Honolulu:IEEE, 2017: 1132-1140.
SAHARIAC, HO J, CHANW, et al. Image super-resolution via iterative refinement[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(4): 4713-4726.
[24]
WANGY, YANGW, CHENX, et al. SinSR: Diffusion-based image super-resolution in a single step[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Seattle:IEEE, 2024: 25796-25805.