Multi-level feature extraction and reconstruction for 3D MRI image super-resolution

作者全名："Li, Hongbi; Jia, Yuanyuan; Zhu, Huazheng; Han, Baoru; Du, Jinglong; Liu, Yanbing"

作者地址："[Li, Hongbi; Jia, Yuanyuan; Han, Baoru; Du, Jinglong; Liu, Yanbing] Chongqing Med Univ, Coll Med Informat, Chongqing 400016, Peoples R China; [Zhu, Huazheng] Chongqing Univ Sci & Technol, Coll Intelligent Technol & Engn, Chongqing 401331, Peoples R China; [Liu, Yanbing] Chongqing Municipal Educ Commiss, Chongqing 400020, Peoples R China"

通信作者："Du, JL; Liu, YB (通讯作者)，Chongqing Med Univ, Coll Med Informat, Chongqing 400016, Peoples R China.; Liu, YB (通讯作者)，Chongqing Municipal Educ Commiss, Chongqing 400020, Peoples R China."

来源：COMPUTERS IN BIOLOGY AND MEDICINE

ESI学科分类：COMPUTER SCIENCE

WOS号：WOS:001202432200001

JCR分区：Q1

影响因子：7

年份：2024

卷号：171

期号：　

开始页：　

结束页：　

文献类型：Article

关键词：Super-resolution reconstruction; MRI image; Multi-level feature; Deep learning

摘要："Magnetic resonance imaging (MRI) is an essential radiology technique in clinical diagnosis, but its spatial resolution may not suffice to meet the growing need for precise diagnosis due to hardware limitations and thicker slice thickness. Therefore, it is crucial to explore suitable methods to increase the resolution of MRI images. Recently, deep learning has yielded many impressive results in MRI image super -resolution (SR) reconstruction. However, current SR networks mainly use convolutions to extract relatively single image features, which may not be optimal for further enhancing the quality of image reconstruction. In this work, we propose a multi -level feature extraction and reconstruction (MFER) method to restore the degraded highresolution details of MRI images. Specifically, to comprehensively extract different types of features, we design the triple -mixed convolution by leveraging the strengths and uniqueness of different filter operations. For the features of each level, we then apply deconvolutions to upsample them separately at the tail of the network, followed by the feature calibration of spatial and channel attention. Besides, we also use a soft cross -scale residual operation to improve the effectiveness of parameter optimization. Experiments on lesion -free and glioma datasets indicate that our method obtains superior quantitative performance and visual effects when compared with state-of-the-art MRI image SR methods."

基金机构："Natural Science Foundation of Chongqing, China [CSTB2023NSCQ-MSX0130, cstc2021jcyi-bshX0168]; National Natural Science Foundation of China [62272074]; Young Project of Science and Technology Research Program of Chongqing Education Commission of China [KJQN202001513, KJQN202101501]; Medical Image Intelligent Analysis and Application Innovation Team of Chongqing Medical University, China [ZSK0102]"

基金资助正文："This work was partly supported by the Natural Science Foundation of Chongqing, China (Grant No. CSTB2023NSCQ-MSX0130, cstc2021jcyi-bshX0168) , the National Natural Science Foundation of China (Grant No. 62272074) , the Young Project of Science and Technology Research Program of Chongqing Education Commission of China (No. KJQN202001513, No. KJQN202101501) , and the Medical Image Intelligent Analysis and Application Innovation Team of Chongqing Medical University, China (NO. ZSK0102) ."