基于3D深度残差网络与级联U-Net的缺血性脑卒中病灶分割算法

doi:10.11772/j.issn.1001-9081.2019040717

计算机应用 ›› 2019, Vol. 39 ›› Issue (11): 3274-3279.DOI: 10.11772/j.issn.1001-9081.2019040717

基于3D深度残差网络与级联U-Net的缺血性脑卒中病灶分割算法

王平¹, 高琛¹, 朱莉¹, 赵俊¹, 张晶¹, 孔维铭²

1. 南昌大学信息工程学院, 南昌 330031;
2. 中国人民解放军第九四医院, 南昌 330031

收稿日期:2019-04-26 修回日期:2019-06-24 出版日期:2019-11-10 发布日期:2019-08-21
通讯作者: 朱莉
作者简介:王平(1958-),女,河南安阳人,教授,主要研究方向:图像处理、模式识别;高琛(1993-),女,江西上饶人,硕士研究生,主要研究方向:图像处理、机器学习;朱莉(1982-),女,江西南昌人,副教授,博士,主要研究方向:图像处理、机器学习;赵俊(1995-),女,河南周口人,硕士研究生,主要研究方向:图像处理、机器学习;张晶(1994-),女,河北邯郸人,硕士研究生,主要研究方向:图像处理、机器学习;孔维铭(1984-),男,山东曹县人,硕士研究生,主要研究方向:医学影像分析。
基金资助:
国家自然科学基金资助项目（61463035）；中国博士后基金资助项目（2016M592117）；江西省杰出青年基金资助项目（2018ACB21038）；江西省科技厅科技支撑计划项目（20151BBG70057）；江西省教育厅科技项目（GJJ14137）。

Segmentation algorithm of ischemic stroke lesion based on 3D deep residual network and cascade U-Net

WANG Ping¹, GAO Chen¹, ZHU Li¹, ZHAO Jun¹, ZHANG Jing¹, KONG Weiming²

1. College of Information Engineering, Nanchang University, Nanchang Jiangxi 330031, China;
2. The 94 th Hospital of the Chinese People's Liberation Army, Nanchang Jiangxi 330031, China

Received:2019-04-26 Revised:2019-06-24 Online:2019-11-10 Published:2019-08-21
Supported by:
This work is partially supported by the National Natural Science Foundation of China (61463035), the China Postdoctoral Science Foundation (2016M592117), the Outstanding Youth Fund Project in Jiangxi Province (2018ACB21038), the Science and Technology Support Project of Jiangxi Science and Technology Department (20151BBG70057), the Science and Technology Project of Jiangxi Education Department (GJJ14137).

摘要/Abstract

摘要： 为了解决人工勾画缺血性脑卒中病灶费时费力且易引入主观差异的问题，提出了一种基于三维（3D）深度残差网络与级联U-Net的自动分割算法。首先，为了有效利用图像的3D上下文信息并改善类不平衡现象，将脑卒中核磁共振图像（MRI）采样成图像块作为网络输入；然后，利用基于3D深度残差网络与级联U-Net的分割模型对图像块进行特征提取，获得粗分割结果；最后，对粗分割结果进行精分割处理。在ISLES数据集上的实验结果表明，该算法的Dice系数可达到0.81，精确度可达到0.81，灵敏度可达到0.81，平均对称表面距离（ASSD）距离系数为1.32，HD为22.67。所提算法与3D U-Net算法、基于水平集算法、基于模糊C均值（FCM）算法和基于卷积神经网络（CNN）算法相比分割性能更好。

关键词: 急性缺血性脑卒中, 自动分割, 三维, 磁共振图像

Abstract: Artificial identification of ischemic stroke lesion is time-consuming, laborious and easy be added subjective differences. To solve this problem, an automatic segmentation algorithm based on 3D deep residual network and cascade U-Net was proposed. Firstly, in order to efficiently utilize 3D contextual information of the image and the solve class imbalance issue, the patches were extracted from the stroke Magnetic Resonance Image (MRI) and put into network. Then, a segmentation model based on 3D deep residual network and cascade U-Net was used to extract features of the image patches, and the coarse segmentation result was obtained. Finally, the fine segmentation process was used to optimize the coarse segmentation result. The experiment results show that, on the dataset of Ischemic Stroke LEsion Segmentation (ISLES), for the proposed algorithm, the Dice similarity coefficient reached 0.81, the recall reached 0.81 and the precision reached 0.81, the distance coefficient Average Symmetric Surface Distance (ASSD) reached 1.32 and Hausdorff Distance (HD) reached 22.67. Compared with 3D U-Net algorithm, level set algorithm, Fuzzy C-Means (FCM) algorithm and Convolutional Neural Network (CNN) algorithm, the proposed algorithm has better segmentation performance.

Key words: acute ischemic stroke, automatic segmentation, three-dimensional, Magnetic Resonance Image (MRI)

中图分类号:

TP391.7

王平, 高琛, 朱莉, 赵俊, 张晶, 孔维铭. 基于3D深度残差网络与级联U-Net的缺血性脑卒中病灶分割算法[J]. 计算机应用, 2019, 39(11): 3274-3279.

WANG Ping, GAO Chen, ZHU Li, ZHAO Jun, ZHANG Jing, KONG Weiming. Segmentation algorithm of ischemic stroke lesion based on 3D deep residual network and cascade U-Net[J]. Journal of Computer Applications, 2019, 39(11): 3274-3279.

参考文献

[1] MOZAFFARIAN D, BENJAMIN E J, GO A S, et al. Heart disease and stroke statistics-2016 update:a report from the American heart association[J]. Circulation, 2016, 133(4):338-360.
[2] ASTRUP J, SYMON L, BRANSTON N M, et al. Cortical evoked potential and extracellular K+ and H+ at critical levels of brain ischemia[J]. Stroke, 1977, 8(1):51-57.
[3] MAIER O, MENZE B H, VON DER GABLENTZ J, et al. ISLES 2015-a public evaluation benchmark for ischemic stroke lesion segmentation from multispectral MRI[J]. Medical Image Analysis, 2017, 35:250-269.
[4] NEETHU S, VENKATARAMAN D. Stroke detection in brain using CT images[C]//Proceedings of the 2014 Artificial Intelligence and Evolutionary Algorithms in Engineering Systems, AISC 324. Berlin:Springer, 2015:379-386.
[5] KARTHIKEYAN S M. EZHILARASI M. Automatic stroke lesion segmentation from diffusion weighted MRI images[J]. International Journal of Engineering Research & Technology, 2016,7(2):111-115.
[6] ELLWAA A, HUSSEIN A, ALNAGGAR E, et al. Brain tumor segmentation using random forest trained on iteratively selected patients[C]//Proceedings of the 2016 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 10154. Berlin:Springer, 2016:129-137.
[7] SONG B, CHOU C, CHEN X, et al. Anatomy-guided brain tumor segmentation and classification[C]//Proceedings of the 2016 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 10154. Berlin:Springer, 2016:162-170.
[8] le FOLGOC L, NORI A V, ANCHA S, et al. Lifted auto-context forests for brain tumor segmentation[C]//Proceedings of the 2016 International Workshop on Brain lesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 10154. Berlin:Springer, 2016:171-183.
[9] CHEN L, BENTLEY P, RUECKERT D. Fully automatic acute ischemic lesion segmentation in DWI using convolutional neural networks[J]. NeuroImage:Clinical, 2017, 15:633-643.
[10] LUCAS C, KEMMLING A, MAMLOUK A M, et al. Multi-scale neural network for automatic segmentation of ischemic strokes on acute perfusion images[C]//Proceedings of the IEEE 15th International Symposium on Biomedical Imaging. Piscataway:IEEE, 2018:1118-1121.
[11] KAMNITSAS K, LEDIG C, NEWCOMBE V F J, et al. Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation[J]. Medical Image Analysis, 2017, 36:61-78.
[12] GUERRERO R, QIN C, OKTAY O, et al. White matter hyperintensity and stroke lesion segmentation and differentiation using convolutional neural networks[J]. NeuroImage:Clinical, 2018, 17:918-934.
[13] HE K, ZHANG X, REN S, et al. Delving deep into rectifiers:surpassing human-level performance on ImageNet classification[C]//Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway:IEEE, 2015:1026-1034.
[14] CICEK Ö, ABDULKADIR A, LIENKAMP S S, et al. 3D U-net:learning dense volumetric segmentation from sparse annotation[C]//Proceedings of the 2016 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 9901. Berlin:Springer, 2016:424-432.
[15] HAECK T, MAES F, SUETENS P. ISLES challenge 2015:automated model-based segmentation of ischemic stroke in MR images[C]//Proceedings of the 2015 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 9556. Berlin:Springer, 2015:246-253.
[16] FENG C, ZHAO D, HUANG M. Segmentation of ischemic stroke lesions in multi-spectral MR images using weighting suppressed FCM and three phase level set[C]//Proceedings of the 2015 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 9556. Berlin:Springer, 2015:233-245.
[17] HAVAEI M, DUTIL F, PAL C, et al. A convolutional neural network approach to brain tumor segmentation[C]//Proceedings of the 2015 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 9556. Berlin:Springer, 2015:195-208.
[18] MCKINLEY R, HÄNI L, WIEST R, et al. Segmenting the ischemic penumbra:a decision forest approach with automatic threshold finding[C]//Proceedings of the 2015 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 9556. Berlin:Springer, 2015:275-283.
[19] MAIER O, WILMS M, HANDELS H. Image features for brain lesion segmentation using random forests[C]//Proceedings of the 2015 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 9556. Berlin:Springer, 2015:119-130.
[20] ROBBEN D, CHRISTIAENS D, RANGARAJAN J, et al. A voxel-wise, cascaded classification approach to ischemic stroke lesion segmentation[C]//Proceedings of the 2015 International Workshop on Brainlesion:Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, LNCS 9556. Berlin:Springer, 2015:254-265.

基于3D深度残差网络与级联U-Net的缺血性脑卒中病灶分割算法

Segmentation algorithm of ischemic stroke lesion based on 3D deep residual network and cascade U-Net

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	高工, 杨红雨, 刘洪. 基于深度学习的三维点云人脸识别[J]. 计算机应用, 2021, 41(9): 2736-2740.
[2]	刘双元, 郑王里, 林云汉. 面向三维特征描述子的自适应二进制简化方法[J]. 计算机应用, 2021, 41(7): 2062-2069.
[3]	黄梨, 卢龙. 基于长距离依赖编码与深度残差U-Net的缺血性卒中病灶分割[J]. 计算机应用, 2021, 41(6): 1820-1827.
[4]	李伟强, 雷航, 张静玉, 王旭鹏. 基于标签分布学习的三维手部姿态估计[J]. 计算机应用, 2021, 41(2): 550-555.
[5]	徐少伟, 秦品乐, 曾建朝, 赵致楷, 高媛. 基于注意力机制的两阶段纵膈淋巴结自动分割算法[J]. 计算机应用, 2021, 41(2): 556-562.
[6]	陈莉, 王洪元, 张云鹏, 曹亮, 殷雨昌. 联合均等采样随机擦除和全局时间特征池化的视频行人重识别方法[J]. 计算机应用, 2021, 41(1): 164-169.
[7]	赵青, 余元辉. 基于分层特征化网络的三维人脸识别[J]. 计算机应用, 2020, 40(9): 2514-2518.
[8]	张豪, 张强, 邵思羽, 丁海斌. 深度学习在单图像三维模型重建的应用[J]. 计算机应用, 2020, 40(8): 2351-2357.
[9]	彭育辉, 郑玮鸿, 张剑锋. 基于深度学习的道路障碍物检测方法[J]. 计算机应用, 2020, 40(8): 2428-2433.
[10]	裴焱栋, 顾克江. 基于内容和语义的三维模型检索综述[J]. 计算机应用, 2020, 40(7): 1863-1872.
[11]	熊勇, 张加, 余嘉俊, 张本任, 梁萱卓, 朱奇舸. 船舶三维管路智能布局优化算法[J]. 计算机应用, 2020, 40(7): 2164-2170.
[12]	曹小威, 曲志坚, 徐玲玲, 刘晓红. 基于胶囊网络的三维模型识别[J]. 计算机应用, 2020, 40(5): 1309-1314.
[13]	屈立成, 吕娇, 赵明, 王海飞, 屈艺华. 基于三维时空地图和运动分解的多机器人路径规划算法[J]. 计算机应用, 2020, 40(12): 3499-3507.
[14]	周健, 黄章进. 基于改进三维形变模型的三维人脸重建和密集人脸对齐方法[J]. 计算机应用, 2020, 40(11): 3306-3313.
[15]	魏小娜, 邢嘉祺, 王振宇, 王颖珊, 石洁, 赵地, 汪红志. 基于改进U-Net的关节滑膜磁共振图像的分割[J]. 计算机应用, 2020, 40(11): 3340-3345.