Attention mechanism combined with convolutional neural network ( CNN ) achieves promising performance for magnetic resonance imaging ( MRI) image segmentation, however these methods on-ly learn attention weights from single scale, resulting in incomplete attention learning. A novel meth-od named completed attention convolutional neural network ( CACNN ) is proposed for MRI image segmentation. Specifically, the channel-wise attention block ( CWAB) and the pixel-wise attention block ( PWAB) are designed to learn attention weights from the aspects of channel and pixel levels. As a result, completed attention weights are obtained, which is beneficial to discriminative feature learning. The method is verified on two widely used datasets (HVSMR and MRBrainS), and the ex-perimental results demonstrate that the proposed method achieves better results than the state-of-the-art methods.

ZHANG Zhong;LV Shijie;LIU Shuang;XIAO Baihua

Tianjin Key Laboratory of Wireless Mobile Communications and Power Transmission,Tianjin Normal University,Tianjin 300387,P.R.China;State Key Laboratory for Management and Control of Complex Systems,Institute of Automation,Chinese Academy of Sciences,Beijing 100190,P.R.China

高技术通讯（英文版）

[1]ZHANG Zhong;LV Shijie;LIU Shuang;XIAO Baihua-.Completed attention convolutional neural network for MRI image segmentation)[J].高技术通讯（英文版）,2022(03):247-251

Completed,CACNN,CWAB,PWAB,HVSMR,MRBrainS

attention,convolutional,neural,network,image,segmentation,Attention,mechanism,combined,achieves,promising,performance,magnetic,resonance,imaging,however,these,methods,weights,from,single,scale,resulting,incomplete,learning,novel,named,completed,proposed,Specifically,channel,wise,block,pixel,are,designed,aspects,levels,obtained,which,beneficial,discriminative,feature,verified,widely,used,datasets,ex,perimental,results,demonstrate,that,better,than,state,art

0.437072