3D CROSS-SCALE FEATURE TRANSFORMER NETWORK FOR BRAIN MR IMAGE SUPER-RESOLUTION

作者全名："Zhang, Wanqi; Wang, Lulu; Chen, Wei; Jia, Yuanyuan; He, Zhongshi; Du, Jinglong"

作者地址："[Zhang, Wanqi; Wang, Lulu; Chen, Wei; He, Zhongshi] Chongqing Univ, Coll Comp Sci, Chongqing, Peoples R China; [Jia, Yuanyuan; Du, Jinglong] Chongqing Med Univ, Coll Med Informat, Chongqing, Peoples R China"

通信作者："He, ZS (通讯作者)，Chongqing Univ, Coll Comp Sci, Chongqing, Peoples R China.; Du, JL (通讯作者)，Chongqing Med Univ, Coll Med Informat, Chongqing, Peoples R China."

来源："2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)"

ESI学科分类：　

WOS号：WOS:000864187901126

JCR分区：　

影响因子：　

年份：2022

卷号：　

期号：　

开始页：1356

结束页：1360

文献类型：Proceedings Paper

关键词：Magnetic resonance image; Cross-scale self-similarity; Super-resolution; Attention mechanism

摘要："High-resolution (HR) magnetic resonance (MR) images could provide reliable visual information for clinical diagnosis. Recently, super-resolution (SR) methods based on convolutional neural networks (CNNs) have shown great potential in obtaining HR MR images. However, most existing CNN-based SR methods neglect the internal priors of the MR image, which hides the performance of SR. In this work, we propose a 3D cross-scale feature transformer network (CFTN) to utilize the cross-scale priors within MR features. Specifically, we stack multiple 3D residual channel attention blocks (RCABs) as the backbone. Meanwhile, we design a plug-in mutual-projection feature enhancement module (MFEM) to extract the target-scale features with HR cues, which is able to capture the global cross-scale self-similarity within features and can be flexibly inserted into any position of the backbone. Furthermore, we propose a spatial attention fusion module (SAFM) to adaptively adjust and fuse the target-scale features and upsampled features that are respectively extracted by the MFEM and the backbone. Experimental results show that our CFTN achieves a new state-of-the-art MR image SR performance."

基金机构："Intelligent Medical Project of Chongqing Medical University [ZHYXQNRC202101]; Graduate Research and Innovation Foundation of Chongqing, China [CYS21060]"

基金资助正文："This work was supported by the Intelligent Medical Project of Chongqing Medical University (ZHYXQNRC202101), the Graduate Research and Innovation Foundation of Chongqing, China (Grant No.CYS21060)."