多模态3D卷积神经网络脑部胶质瘤分割方法

详细信息    查看全文 | 推荐本文 |

英文篇名：Multimodal 3D Convolutional Neural Networks for Brain Glioma Segmentation 作者：谷宇 ; 吕晓琪 ; 李菁 ; 任国印 ; 喻大华 ; 赵瑛 ; 吴凉 ; 张文莉 ; 郝小静 ; 黄显武 英文作者：GU Yu;Lü Xiao-qi;LI Jing;REN Guo-yin;YU Da-hua;ZHAO Ying;WU Liang;ZHANG Wen-li;HAO Xiao-jing;HUANG Xian-wu;School of Computer Engineering and Science,Shanghai University;Inner Mongolia Key Laboratory of Pattern Recognition and Intelligent Image Processing,School of Information Engineering,Inner Mongolia University of Science and Technology; 关键词：脑部胶质瘤 ; 瘤内结构 ; 多模态MRI ; 3D卷积神经网络 ; 图像分割 英文关键词：brain glioma;;intra-tumoral structures;;multimodality MRI;;3D convolution neural network;;image segmentation 中文刊名：KXJS 英文刊名：Science Technology and Engineering 机构：上海大学计算机工程与科学学院;内蒙古科技大学信息工程学院内蒙古自治区模式识别与智能图像处理重点实验室; 出版日期：2018-03-08 出版单位：科学技术与工程 年：2018 期：v.18;No.440 基金：国家自然科学基金(61771266,81571753,81460279,81301281,61179019);; 内蒙古自治区自然科学基金(2015MS0604,2014MS0828);; 内蒙古自治区高等学校科学研究项目(NJZY145,NJZZ14161);; 包头市科技计划项目(2015C2006-14);; 内蒙古科技大学创新基金(2015QNGG03,2014QNGG08,2015QDL26,2014QDL045)资助 语种：中文; 页：KXJS201807004 页数：7 CN：07 ISSN：11-4688/T 分类号：23-29

摘要

由于大多数脑部胶质瘤边界有水肿且内部结构复杂,分割胶质瘤及瘤内结构难度较大。提出一种新的基于多模态MRI 3D卷积神经网络(CNN)脑部胶质瘤及瘤内各结构的自动分割算法。首先,标准化由T1、T1c、T2、FLAIR 4个MRI模态组成的输入图像。其次,构建10个卷积层,2个全连接层的3D CNN。卷积层采用3×3×3的3D卷积核;全连接层采用PRe Lu激励函数,并结合dropout技术防止过拟合。构建的3D CNN分割胶质瘤和瘤内各结构精度高,与专家手动分割的结果接近。实验结果表明,构建的多模态3D CNN能够准确地分割MRI多模态图像脑部胶质瘤及瘤内各结构,具有重要的临床意义。
        Brain glioma is difficult to be segmented as existing edema regions and its complex intra-tumoral structures. A novel algorithm about multimodality 3 D convolution neural network( CNN) was proposed to segment brain glioma and its internal structures automatically. The input images were normalized composed of T1,T1 c,T2,FLAIR 4 MRI modalities. Then,the 3 D CNN consisted of 10 convolution layers and 2 fully connected layers. The convolution layers used 3 × 3 × 3 3 D convolution kernels,and the full connection layers used PRe Lu activation function,combined with dropout technique to prevent over-fitting. The constructed 3 D CNN segmented gliomas and intratumoral structures with high accuracy,which close to the ground truth provided by experts. The experimental results showed that the constructed MRI multimodality 3 D CNN could segment the brain gliomas and its intra-tumoral structures accurately,which is of great clinical significance.

引文

1罗蔓,黄靖,杨丰.基于多模态3D-CNNs特征提取的MRI脑肿瘤分割方法.科学技术与工程,2014;14(31):78-83Luo Man,Huang Jing,Yang Feng.Multimodal 3D convolutional neural networks features for brain tumor segmentation.Science Technology and Engineering,2014;14(31):78-83
    2 Stadl bauer A,Moser E,Gruber S,et al.Improved delineation of brain tumors:an automated method for segmentation based on pathologic changes of 1H-MRSI metabolites in gliomas.Neuroimage,2004;23(2):454-461
    3 Shanthi K J,Kumar M S.Skull stripping and automatic segmentation of brain MRI using seed growth and threshold techniques.proceedings of the International Conference on Intelligent and Advanced Systems,Malaysia,F 2007.11.25,2007.New York:IEEE Computer Society,2007
    4 Chong V F,Zhou J Y,Khoo J B,et al.Tongue carcinoma:tumor volume measurement.International Journal of Radiation Oncology Biology Physics,2004;59(1):59-66
    5张雯,王小鹏,李志强,等.分水岭优化的C-V模型脑肿瘤图像分割.计算机工程与应用,2017;53(5):176-180Zhang Wen,Wang Xiaopeng,Li Zhiqiang,et al.Brain tumor image segmentation based on C-V model optimized by watershed transformation.Computer Engineering and Applications,2017;53(5):176-180
    6林相波,王新宁,郭冬梅.一种分割脑磁共振图像的改进FCM聚类算法.中国生物医学工程学报,2016;35(6):648-657Lin Xiangbo,Wang Xinning,Guo Dongmei.A modified FCM clustering method for brain magnetic resonance image segmentation.Chinese Journal of Biomedical Engineering,2016;35(6):648-657
    7宋艳涛,纪则轩,孙权森.基于图像片马尔科夫随机场的脑MR图像分割算法.自动化学报,2014;40(8):1754-1763Song Yantao,Ji Zexuan,Sun Quansen.Brain MR image segmentation algorithm based on Markov random field with image patch.Acta Automatica Sinica,2014;40(8):1754-1763
    8曹家梓,宋爱国.基于马尔科夫随机场的纹理图像分割方法研究.仪器仪表学报,2015;36(4):776-786Cao Jiazi,Song Aiguo.Research on the texture image segmentation method based on Markov random field.Chinese Journal of Scientific Instrument,2015;36(4):776-786
    9 Vijayakumar C,Damayanti G,Pant R,et al.Segmentation and grading of brain tumors on apparent diffusion coefficient images using selforganizing maps.Computerized Medical Imaging&Graphics,2007;31(7):473-484
    10 Khotanlou H,Colliot O,Atif J,et al.3D brain tumor segmentation in MRI using fuzzy classification,symmetry analysis and spatially constrained deformable models.Fuzzy Sets&Systems,2009;160(10):1457-1473
    11张腾达,吕晓琪,任晓颖,等.基于模糊水平集的脑肿瘤MR图像分割方法.现代电子技术,2016;39(18):91-95Zhang Tengda,Lu Xiaoqi,Ren Xiaoying,et al.Brain tumor MRimage segmentation method based on fuzzy level set.Modern Electronics Technique,2016;39(18):91-95
    12 Zikic D,Ioannou Y,Brown M,et al.Segmentation of brain tumor tissues with convolutional neural Networks.Proceedings of the Miccai Workshop on Multimodal Brain Tumor Segmentation Challenge,United States,F,2014.Berlin:Springer,2014
    13周飞燕,金林鹏,董军.卷积神经网络研究综述.计算机学报,2017;40(7):1-23Zhou Feiyan,Jin Linpeng,Dong Jun,Review of convolutional neural network.Chinese Journal of Computers,2017;40(7):1-23
    14罗蔓.结合MRI多模态信息与3D-CNNs特征提取的脑肿瘤分割研究.广州:南方医科大学,2015Luo Man.The research of brain tumor segmentation based on MRImulti-modality images and 3D-CNNs features.Guangzhou:Southern Medical University,2015
    15 Menze B H,Jakab A,Bauer S,et al.The multimodal brain tumor image segmentation benchmark(BRATS).IEEE Transactions on Medical Imaging,2015;34(10):1993-2024
    16 Kamnitsas K,Ledig C,Newcombe V F,et al.Efficient multi-scale3D CNN with fully connected CRF for accurate brain lesion segmentation.Medical Image Analysis,2017;36:61-78
    17 Chen X,Nguyen B P,Chui C K,et al.Reworking multilabel brain tumor segmentation:An automated framework using structured kernel sparse representation.IEEE Systems,Man,and Cybernetics Magazine,2017;3(2):18-22
    18 Zikic D,Glocker B,Konukoglu E,et al.Context-sensitive classification forests for segmentation of brain tumor tissues.Proceedings of the MICCAI 2012 Challenge on Multimodal Brain Tumor Segmentation(Bra TS),France,F,2012.Berlin:Springer