Automatic segmentation of ischemic stroke lesions from computed tomography(CT)images is of great significance for identifying and curing this life-threatening condition.However,in addition to the problem of low image contrast,it is also challenged by the complex changes in the appearance of the stroke area and the difficulty in obtaining image data.Considering that it is difficult to obtain stroke data and labels,a data enhancement algorithm for one-shot medical image segmentation based on data augmentation using learned transformation was proposed to increase the number of data sets for more accurate segmentation.A deep convolutional neural network based algorithm for stroke lesion segmentation,called structural similarity with light U-structure(USSL)Net,was proposed.We embedded a convolution module that combines switchable normalization,multi-scale convolution and dilated convolution in the network for better segmentation performance.Besides,considering the strong structural similarity between multi-modal stroke CT images,the USSL Net uses the correlation maximized structural similarity loss(SSL)function as the loss function to learn the varying shapes of the lesions.The experimental results show that our framework has achieved results in the following aspects.First,the data obtained by adding our data enhancement algorithm is better than the data directly segmented from the multi-modal image.Second,the performance of our network model is better than that of other models for stroke segmentation tasks.Third,the way SSL functioned as a loss function is more helpful to the improvement of segmentation accuracy than the cross-entropy loss function.

JIANG Zhiguo;CHANG Qing

School of Information Science and Engineering,East China University of Science and Technology,Shanghai 200237,China

上海交通大学学报（英文版）

[1]JIANG Zhiguo;CHANG Qing-.USSL Net:Focusing on Structural Similarity with Light U-Structure for Stroke Lesion Segmentation)[J].上海交通大学学报（英文版）,2022(04):485-497

USSL,Net,Focusing,Structural,Similarity,Light,Structure,Stroke,Lesion,Segmentation,Automatic,segmentation,ischemic,stroke,lesions,from,computed,tomography,images,great,significance,identifying,curing,this,life,threatening,condition,However,addition,problem,contrast,also,challenged,by,complex,changes,appearance,area,difficulty,obtaining,data,Considering,that,labels,enhancement,algorithm,shot,medical,augmentation,learned,transformation,was,proposed,increase,number,sets,more,accurate,deep,convolutional,neural,network,called,structural,similarity,light,structure,We,embedded,module,combines,switchable,normalization,multi,scale,dilated,better,performance,Besides,considering,strong,between,modal,uses,correlation,maximized,loss,varying,shapes,experimental,results,show,our,framework,has,achieved,following,aspects,First,obtained,adding,than,directly,segmented,Second,other,models,tasks,Third,way,functioned,helpful,improvement,accuracy,cross,entropy

0.522726