基于形变场灰度场插值网络的切片间超分辨率

doi:10.3969/j.issn.1006-2475.2024.03.011

摘要/Abstract

摘要：
摘要：磁共振成像是一种常见的医学影像方法。由于硬件等因素的限制，MR图像切片间分辨率会远低于切片内分辨率，图像的质量较低，影响医生对患者病情的诊断，因此需要提高切片间分辨率，以便显示图像更多的细节。为了实现切片间超分辨率，提出一种基于非参数化形变场和灰度场插值网络的算法。首先，根据图像配准原理，利用近似U-Net的网络对低分辨率图像中2张相邻切片进行无监督训练，以生成切片间的双向形变场。然后，利用第一张切片与形变场生成配准后的图像，将配准后的图像与第二张切片进行训练，得到它们之间的双向灰度场，由此得到相邻2张切片中间任意位置的形变场和灰度场。最后，利用插值的方法获得中间的任一位置切片。与其他现有的算法相比，该算法在视觉效果和客观评价指标上都有显著提升，其中PSNR和SSIM分别均在30 dB和0.99以上。

关键词: 关键词：磁共振成像, 超分辨率, 形变场, 灰度场

Abstract:

Abstract： Magnetic resonance imaging is a widely used medical imaging method. Constrained by hardware conditions and other factors， the inter-slice resolution of MR images is much lower than the intra-slice resolution， resulting in low image quality and affecting the doctor’s diagnosis of the patient’s condition. Therefore， it is necessary to improve the inter-slice resolution to display more details of the image. In order to achieve super resolution between slices， an algorithm based on non-parametric deformation field and gray field interpolation network is proposed. Firstly， according to the principle of image registration， an approximate U-Net network is used for unsupervised training of two adjacent slices in low-resolution images to generate bidirectional deformation fields between slices. Then， the registered image is generated by using the first slice and the deformation field， and the registered image and the second slice are trained to obtain the bidirectional gray field between them， from which the deformation field and gray field at any position in the middle of the adjacent two slices can be obtained. Finally， any position slice in the middle can be obtained by interpolation method. Compared to other existing algorithms， this algorithm has significant improvement in visual effect and objective evaluation index， with PSNR and SSIM above 30 dB and 0.99， respectively.

Key words: Key words： magnetic resonance imaging, super-resolution, deformation field, grayscale field

中图分类号:

TP391

刘迅, 张东, 袁达龙. 基于形变场灰度场插值网络的切片间超分辨率[J]. 计算机与现代化, 2024, 0(03): 67-71.

LIU Xun, ZHANG Dong, YUNG Da-long. Inter-slice Super-resolution Based on Deformation Field and Grey Field Interpolation Networks[J]. Computer and Modernization, 2024, 0(03): 67-71.

参考文献

［1］ AL-GALAL S A Y， ALSHAIKHLI I F T， ABDULRAZZAQ
M M. MRI brain tumor medical images analysis using deep learning techniques： A systematic review［J］. Health and Technology， 2021，11（2）：267-282.
［2］ GREENSPAN H， OZ G， KIRYATI N， et al. MRI inter-slice reconstruction using super-resolution［J］. Magnetic Resonance Imaging， 2002，20（5）：437-446.
［3］曹泽红，刘高平，张志强，等. 基于多模态MRI脑影像的超分辨率重建［J］. 南方医科大学学报， 2022，42（7）：1019-1025.
［4］ CARMI E， LIU S Y， ALON N， et al. Resolution enhancement in MRI［J］. Magnetic Resonance Imaging， 2006，24（2）：133-154.
［5］ CHAUDHARI A S， FANG Z N， KOGAN F， et al. Super-resolution musculoskeletal MRI using deep learning［J］. Magnetic Resonance in Medicine， 2018，80（5）：2139-2154.
［6］ LEDIG C， THEIS L， HUSZAR F， et al. Photo-realistic single image super-resolution using a generative adversarial network［C］// Proceedings of the 30th IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. 2017：105-114.
［7］ ROUSSEAU F， KIM K， STUDHOLME C， et al. On super-resolution for fetal brain MRI［C］// Proceedings of the 13th International Conference on Medical Image Computing and Computer-Assisted Intervention （MICCAI）. 2010：355-362.
［8］ SUI Y， AFACAN O， GHOLIPOUR A， et al. MRI super-resolution through generative degradation learning［C］// Proceedings of the International Conference on Medical Image Computing and Computer Assisted Intervention （MICC-
AI）. 2021：430-440.
［9］ TOLEDO N V， RUEDA A， MARTA C S， et al. Super-resolution in cardiac MRI using a Bayesian approach［C］// Proceedings of the Conference on Medical Imaging. 2013：1-6.
［10］ WANG H， WU W， SU Y， et al. Image super-resolution using a improved generative adversarial network［C］// Proceedings of the 9th IEEE International Conference on Electronics Information and Emergency Communication （ICEIE-
C）. 2019.
［11］ ZHANG H T， SHINOMIYA Y， YOSHIDA S， et al. 3D brain MRI reconstruction based on 2D super-resolution technology［C］// Proceedings of the IEEE International Conference on Systems， Man， and Cybernetics （SMC）. 2020：18-23.
［12］ ZHANG K， HU H J， PHILBRICK K， et al. SOUP-GAN： Super-resolution MRI using generative adversarial networks［J］. Tomography， 2022，8（2）：905-919.
［13］ WANG L L， DU J L， ZHU H Z， et al. Brain MR image super-resolution using 3D feature attention network［C］// Proceedings of the IEEE International Conference on Bioinformatics and Biomedicine （IEEE BIBM）. 2020：1151-1155.
［14］ ZHENG Y J， ZHEN B W， CHEN A C， et al. A hybrid convolutional neural network for super-resolution reconstruction of MR images［J］. Medical Physics， 2020，47（7）：3013-3022.
［15］李萌，秦品乐，曾建潮，等. 基于磁共振影像层间插值的超分辨率及多视角融合［J］. 计算机应用， 2021，41（11）：3362-3367.
［16］ POOT D H J， VAN MEIR V， SIJBERS J. General and efficient super-resolution method for multi-slice MRI［C］// Proceedings of the 13th International Conference on Medical Image Computing and Computer-Assisted Intervention （MICCAI）. 2010：862-870.
［17］ PHAM C H， DUCOURNAU A， FABLET R， et al. Brain MRI super-resolution using deep 3D convolutional networks［C］// Proceedings of the IEEE 14th International Symposium on Biomedical Imaging （ISBI）. 2017. DOI：10.
1016/j.compmedimag.2019.101647.
［18］ LEE H， SONG J. Introduction to convolutional neural network using Keras； an understanding from a statistician［J］. Communications for Statistical Applications and Methods， 2019，26（6）：591-610.
［19］ LIU J， CHEN F， WANG X Y， et al. An edge enhanced srgan for MRI super resolution in slice-selection direction［M］// Multimodal Brain Image Analysis and Mathematical Foundations of Computational Anatomy. 2019：12-20.
［20］ BALAKRISHNAN G， ZHAO A， SABUNCU M R， et al. VoxelMorph： A learning framework for deformable medical image registration［J］. IEEE Transactions on Medical Imaging， 2019，38（8）：1788-1800.
［21］ RONNEBERGER O， FISCHER P， BROX T. U-Net： Convolutional networks for biomedical image segmentation［M］// Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. 2015：234-241.
［22］ XU J， LI Z S， DU B W， et al. Reluplex made more practical： Leaky ReLU［C］ // Proceedings of the 25th IEEE Symposium on Computers and Communications （ISCC）. 2020：1-7.
［23］ NI K S， NGUYEN T Q， IEEE. Adaptable K-nearest neighbor for image interpolation［C］// Proceedings of the 33rd IEEE International Conference on Acoustics， Speech and Signal Processing. 2008：1297-1300.
［24］ MASTYLO M. On interpolation of bilinear operators［J］. Journal of Functional Analysis， 2004，214（2）：260-283.
［25］ ZHAO H， GALLO O， FROSIO I， et al. Loss functions for image restoration with neural networks［J］. IEEE Transactions on Computational Imaging， 2017，3（1）：47-57.
［26］ AVANTS B B， EPSTEIN C L， GROSSMAN M， et al. Symmetric diffeomorphic image registration with cross-correlation： Evaluating automated labeling of elderly and neurodegenerative brain［J］. Medical Image Analysis， 2008，
12（1）：26-41.
［27］ OMARA A N， SALEM T M， ELSANADILY S， et al. SSIM-based sparse image restoration［J］. Journal of King Saud University-Computer and Information Sciences， 2022， 34（8）：6243-6254.

[1]	王鑫, 余磊. 多尺度双注意力的图像超分辨率重建方法[J]. 计算机与现代化, 2024, 0(08): 77-87.
[2]	邢世帅, 刘丹凤, 王立国, 潘月涛, 孟灵鸿, 岳晓晗. 基于空间注意力残差网络的图像超分辨率重建模型[J]. 计算机与现代化, 2023, 0(10): 45-52.
[3]	吴丽君, 蔡周威, 陈志聪. 基于改进残差特征蒸馏的轻量级超分辨率网络[J]. 计算机与现代化, 2022, 0(11): 89-94.
[4]	陈彤, 周登文. 基于多级Transformer的超大倍率重建网络:参考图像超分辨率[J]. 计算机与现代化, 2022, 0(08): 121-126.
[5]	金安安, 丁双双, 熊卿智, 许婷婷, 习淑萍, 马继忠. 基于SWOMP算法的EPMA影像优化重构[J]. 计算机与现代化, 2020, 0(12): 104-111.
[6]	梁超, 黄洪全. 基于卷积神经网络的轻量级图像超分辨率[J]. 计算机与现代化, 2020, 0(11): 23-27.
[7]	刘亚男，彭仁勇，王琳. 基于自适应分数阶全变分的超分辨率图像重建[J]. 计算机与现代化, 2018, 0(09): 56-.
[8]	张颢1,2,范新南1,2,3,李敏1,3,张学武1,3. 基于光学成像模型的水下图像超分辨率重构[J]. 计算机与现代化, 2017, 0(4): 7-13.
[9]	解虹1，2，李展1，2，陈慧琼1，2，郑嘉伟1，2，温力1，2. 基于超分辨率的地基观测图像增强系统[J]. 计算机与现代化, 2016, 251(07): 1-5,12.
[10]	李伟祥,李展,吴少春,邓福明. 基于超分重建技术的图像增强系统[J]. 计算机与现代化, 2014, 0(5): 94-98.