高级搜索

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

一种在MR图像中进行脑胶质瘤检测和病灶分割的方法

陈皓 李广 刘洋 强永乾

陈皓, 李广, 刘洋, 强永乾. 一种在MR图像中进行脑胶质瘤检测和病灶分割的方法[J]. 电子与信息学报. doi: 10.11999/JEIT200033
引用本文: 陈皓, 李广, 刘洋, 强永乾. 一种在MR图像中进行脑胶质瘤检测和病灶分割的方法[J]. 电子与信息学报. doi: 10.11999/JEIT200033
Hao CHEN, Guang LI, Yang LIU, Yongqian QIANG. A Glioma Detection and Segmentation Method in MR Imaging[J]. Journal of Electronics and Information Technology. doi: 10.11999/JEIT200033
Citation: Hao CHEN, Guang LI, Yang LIU, Yongqian QIANG. A Glioma Detection and Segmentation Method in MR Imaging[J]. Journal of Electronics and Information Technology. doi: 10.11999/JEIT200033

一种在MR图像中进行脑胶质瘤检测和病灶分割的方法

doi: 10.11999/JEIT200033
基金项目: 国家自然科学基金(61876138, 61203311),陕西省自然科学基金面上项目(2019JM-365),陕西省教育厅自然科学专项(17JK0701),西安邮电大学研究生创新基金(CXJJ2017036)
详细信息
    作者简介:

    陈皓:男,1978年生,博士,副教授,硕士研究生导师,主要研究领域为医疗大数据

    李广:男,1995年生,硕士生,研究方向为计算智能与数据挖掘

    刘洋:男,1995年生,硕士生,研究方向为计算智能与数据挖掘

    强永乾:男,1965年生,博士,副教授,硕士研究生导师,研究方向为医学影像学

    通讯作者:

    陈皓 chenhao@xupt.edu.cn

  • 中图分类号: TP391.41, R445.2

A Glioma Detection and Segmentation Method in MR Imaging

Funds: The National Nature Science Foundation of China (61876138, 61203311), The Natural Science Basic Research Program of Shaanxi Province (2019JM-365), The Scientific Research Program Funded by Shaanxi Provincial Education Department of China (17JK0701), The Graduate Innovation Foundation of Xi’an University of Posts & Telecommunications (CXJJ2017036)
  • 摘要: 针对磁共振图像(MRI)进行脑胶质瘤检测及病灶分割对临床治疗方案的选择和手术实施过程的引导都有着重要的价值。为了提高脑胶质瘤的检测效率和分割准确率,该文提出了一种两阶段计算方法。首先,设计了一个轻量级的卷积神经网络,并通过该网络完成MR图像中肿瘤的快速检测及大致定位;接着,通过集成学习过程对肿瘤周围水肿、肿瘤非增强区、肿瘤增强区和正常脑组织等4种不同区域进行分类和彼此边界的精细分割。为提高分割的准确率,在MR图像中提取了416维影像组学特征并与128维通过卷积神经网络提取的高阶特征进行组合和特征约简,将特征约简后产生的298维特征向量用于分类学习。为对算法的性能进行验证,在BraTS2017数据集上进行了实验,实验结果显示该文提出的方法能够快速检测并定位肿瘤,同时相比其它方法,整体分割精度也有明显提升。
  • 图  1  两阶段计算方法

    图  2  LocNet模型结构

    图  3  CNN特征提取

    图  4  特征的多层次融合约简

    图  5  病灶组织分类与边界分割

    图  6  典型病例分割结果

    图  7  分割结果对比图

    图  8  3组特征集合聚类结果

    图  9  不同特征数量对分类准确率的影响

    图  10  网格大小对定位结果的影响

    图  11  邻域大小对精细化分割的影响

    表  1  实验统计结果

    方法Dice系数灵敏度特异性
    WTTCETWTTCETWTTCET
    均值0.8820.8460.8020.9220.9040.8790.9930.9930.998
    标准差0.0550.0840.1210.0690.0730.0630.0100.0060.002
    中值0.9040.8450.7950.9380.9320.8750.9960.9940.999
    第1四分位数0.8630.7990.7650.8850.8410.8290.9930.9910.997
    第3四分位数0.9380.8960.8650.9810.9670.9190.9980.9970.999
    下载: 导出CSV

    表  2  实验结果对比

    方法Dice系数灵敏度特异性
    WTTCETWTTCETWTTCET
    本文方法0.8820.8460.8020.9220.9040.8790.9930.9930.998
    文献[22]0.8680.7380.6490.8880.7580.7770.9920.9960.9972
    文献[23]0.8970.8250.7640.9120.8410.7750.9940.9970.999
    文献[24]0.9090.8660.7110.8970.8310.7710.9950.9980.998
    文献[25]0.8540.7080.7220.9280.7660.7540.9860.9940.997
    下载: 导出CSV

    表  3  实验效率对比

    Dice-WT Dice-TC Dice-ET 时间(s)
    本文方法 0.882 0.846 0.802 258.3
    RF 0.854 0.802 0.770 796.5
    XGBoost 0.882 0.845 0.802 814.6
    U-Net 0.796 0.769 0.681 4.2
    下载: 导出CSV

    表  4  3组特征聚类结果的评价结果

    方法AMI均一性V-Measure
    CNN0.43610.43630.4657
    Radiomics0.29720.29740.3105
    CNN + Radiomics0.47960.47980.4816
    下载: 导出CSV
  • [1] FURNARI F B, FENTON T, BACHOO R M, et al. Malignant astrocytic glioma: Genetics, biology, and paths to treatment[J]. Genes & Development, 2007, 21(21): 2683–2710. doi:  10.1101/gad.1596707
    [2] OHGAKI H and KLEIHUES P. Population-based studies on incidence, survival rates, and genetic alterations in astrocytic and oligodendroglial gliomas[J]. Journal of Neuropathology & Experimental Neurology, 2005, 64(6): 479–489.
    [3] SACHDEVA J, KUMAR V, GUPTA I, et al. Segmentation, feature extraction, and multiclass brain tumor classification[J]. Journal of Digital Imaging, 2013, 26(6): 1141–1150. doi:  10.1007/s10278-013-9600-0
    [4] SOLTANINEJAD M, YANG Guang, LAMBROU T, et al. Automated brain tumour detection and segmentation using superpixel-based extremely randomized trees in FLAIR MRI[J]. International Journal of Computer Assisted Radiology and Surgery, 2017, 12(2): 183–203. doi:  10.1007/s11548-016-1483-3
    [5] SONG Guoli, HUANG Zheng, ZHAO Yiwen, et al. A Noninvasive system for the automatic detection of gliomas based on hybrid features and PSO-KSVM[J]. IEEE Access, 2019, 7: 13842–13855. doi:  10.1109/ACCESS.2019.2894435
    [6] CAO Zhantao, DUAN Lixin, YANG Guowu, et al. Breast tumor detection in ultrasound images using deep learning[C]. The 3rd International Workshop on Patch-based Techniques in Medical Imaging, Quebec City, Canada, 2017: 121–128.
    [7] SHKOLYAR E, JIA Xiao, CHANG T C, et al. Augmented bladder tumor detection using deep learning[J]. European Urology, 2019, 76(6): 714–718. doi:  10.1016/j.eururo.2019.08.032
    [8] ÖZYURT F, SERT E, and AVCI D. An expert system for brain tumor detection: Fuzzy c-means with super resolution and convolutional neural network with extreme learning machine[J]. Medical Hypotheses, 2019, 134: 109433. doi:  10.1016/j.mehy.2019.109433
    [9] KUMAR S, NEGI A, SINGH J H, et al. Brain tumor segmentation and classification using MRI images via fully convolution neural networks[C]. 2018 International Conference on Advances in Computing, Communication Control and Networking, Greater Noida, India, 2018: 1178–1181.
    [10] ZHAO Xiaomei, WU Yihong, SONG Guidong, et al. A deep learning model integrating FCNNs and CRFs for brain tumor segmentation[J]. Medical Image Analysis, 2018, 43: 98–111. doi:  10.1016/j.media.2017.10.002
    [11] YANG Tiejun and SONG Jikun. An automatic brain tumor image segmentation method based on the U-Net[C]. 2018 4th International Conference on Computer and Communications, Chengdu, China, 2018: 1600–1604.
    [12] AMIRI A, MAHJOUB M A, and REKIK I. Tree-based ensemble classifier learning for automatic brain glioma segmentation[J]. Neurocomputing, 2018, 313: 135–142. doi:  10.1016/j.neucom.2018.05.112
    [13] MUDGAL T K, GUPTA A, JAIN S, et al. Automated system for brain tumour detection and classification using eXtreme gradient boosted decision trees[C]. 2017 International Conference on Soft Computing and its Engineering Applications, Changa, India, 2017: 1–6.
    [14] 陈忠辉, 凌献尧, 冯心欣, 等. 基于模糊c均值聚类和随机森林的短时交通状态预测方法[J]. 电子与信息学报, 2018, 40(8): 1879–1886. doi:  10.11999/JEIT171090

    CHEN Zhonghui, LING Xianyao, FENG Xinxin, et al. Short-term traffic state prediction approach based on FCM and random forest[J]. Journal of Electronics &Information Technology, 2018, 40(8): 1879–1886. doi:  10.11999/JEIT171090
    [15] RAJAGOPAL R. Glioma brain tumor detection and segmentation using weighting random forest classifier with optimized ant colony features[J]. International Journal of Imaging Systems and Technology, 2019, 29(3): 353–359. doi:  10.1002/ima.22331
    [16] CHO H H and PARK H. Classification of low-grade and high-grade glioma using multi-modal image radiomics features[C]. The 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Seogwipo, South Korea, 2017: 3081–3084.
    [17] CHADDAD A, DANIEL P, DESROSIERS C, et al. Novel radiomic features based on joint intensity matrices for predicting glioblastoma patient survival time[J]. IEEE Journal of Biomedical and Health Informatics, 2019, 23(2): 795–804. doi:  10.1109/JBHI.2018.2825027
    [18] 罗会兰, 卢飞, 孔繁胜. 基于区域与深度残差网络的图像语义分割[J]. 电子与信息学报, 2019, 41(11): 2777–2786. doi:  10.11999/JEIT190056

    LUO Huilan, LU Fei, and KONG Fansheng. Image semantic segmentation based on region and deep residual network[J]. Journal of Electronics &Information Technology, 2019, 41(11): 2777–2786. doi:  10.11999/JEIT190056
    [19] KIDO S, HIRANO Y and HASHIMOTO N. Detection and classification of lung abnormalities by use of Convolutional Neural Network (CNN) and regions with cnn features (R-CNN)[C]. 2018 International Workshop on Advanced Image Technology, Chiang Mai, Thailand, 2018: 1–4.
    [20] SHEN Wei, ZHOU Mu, YANG Feng, et al. Multi-scale convolutional neural networks for lung nodule classification[C]. The 24th International Conference on Information Processing in medical Imaging, Sabhal Mor Ostaig, Isle of Skye, UK, 2015: 588–599.
    [21] MENZE B H, JAKAB A, BAUER S, et al. The multimodal brain tumor image segmentation benchmark (BRATS)[J]. IEEE Transactions on Medical Imaging, 2015, 34(10): 1993–2014. doi:  10.1109/TMI.2014.2377694
    [22] CHEN Shengcong, DING Changxing, and ZHOU Chenhong. Brain tumor segmentation with label distribution learning and multi-level feature representation[C]. International Conference on Medical Image Computing and Computer-Assisted Interventions, Quebec, Canada, 2017: 50–53.
    [23] WANG Guotai, LI Wenqi, OURSELIN S, et al. Automatic brain tumor segmentation using cascaded anisotropic convolutional neural networks[C]. The 3rd International MICCAI Brainlesion Workshop, Quebec City, Canada, 2018: 178–190.
    [24] ISLAM M and REN Hongliang. Fully convolutional network with hypercolumn features for brain tumor segmentation[C]. International Conference on Medical Image Computing and Computer-Assisted Interventions, Quebec, Canada, 2017: 108–115.
    [25] ZHOU Fan, LI Tengfei, LI Heng, et al. TP-CNN: A two-phase convolution neural network based model to do automatic brain tumor segmentation by using BRATS 2017 data[C]. International Conference on Medical Image Computing and Computer-Assisted Interventions, Quebec, Canada, 2017: 334–341.
  • [1] 夏平凡, 倪志伟, 朱旭辉, 倪丽萍.  基于双错测度的极限学习机选择性集成方法, 电子与信息学报. 2020, 42(0): 1-9. doi: 10.11999/JEIT190617
    [2] 毛亮, 薛月菊, 魏颖慧, 朱婷婷.  一种用于细粒度人脸识别的眼镜去除方法, 电子与信息学报. 2020, 42(0): 1-9. doi: 10.11999/JEIT200176
    [3] 林得富, 王骏, 蒋亦樟, 王士同.  面向高维数据的Takagi-Sugeno模糊系统建模新方法, 电子与信息学报. 2018, 40(6): 1404-1411. doi: 10.11999/JEIT170792
    [4] 谢涛, 吴恩斯.  一种鲁棒的基于集成学习的核相关红外目标跟踪算法, 电子与信息学报. 2018, 40(3): 602-609. doi: 10.11999/JEIT170527
    [5] 王昌海, 许昱玮, 张建忠.  基于层次分类的手机位置无关的动作识别, 电子与信息学报. 2017, 39(1): 191-197. doi: 10.11999/JEIT160253
    [6] 文伟, 曹雪菲, 张学峰, 陈渤, 王英华, 刘宏伟.  一种基于多极化散射机理的极化SAR图像舰船目标检测方法, 电子与信息学报. 2017, 39(1): 103-109. doi: 10.11999/JEIT160204
    [7] 杜兰, 史蕙若, 李林森, 孙永光, 胡靖.  基于分数阶傅里叶变换的窄带雷达飞机目标回波特征提取方法, 电子与信息学报. 2016, 38(12): 3093-3099. doi: 10.11999/JEIT161035
    [8] 范学满, 胡生亮, 贺静波.  对海雷达目标识别中全极化HRRP的特征提取与选择, 电子与信息学报. 2016, 38(12): 3261-3268. doi: 10.11999/JEIT160722
    [9] 刘家辰, 苗启广, 曹莹, 宋建锋, 权义宁.  基于混合多样性生成与修剪的集成单类分类算法, 电子与信息学报. 2015, 37(2): 386-393. doi: 10.11999/JEIT140161
    [10] 陈博, 王爽, 焦李成, 刘芳, 毛莎莎, 张爽.  利用0-1矩阵分解集成的极化SAR图像分类, 电子与信息学报. 2015, 37(6): 1495-1501. doi: 10.11999/JEIT141059
    [11] 杨兴明, 吴克伟, 孙永宣, 谢昭.  可迁移测度准则下的协变量偏移修正多源集成方法, 电子与信息学报. 2015, 37(12): 2913-2920. doi: 10.11999/JEIT150323
    [12] 陈志锋, 李清宝, 张平, 冯培钧.  基于聚类分析的内核恶意软件特征选择, 电子与信息学报. 2015, 37(12): 2821-2829. doi: 10.11999/JEIT150387
    [13] 徐涛, 杨奇川, 吕宗磊.  一种基于动态集成学习的机场噪声预测模型, 电子与信息学报. 2014, 36(7): 1631-1636. doi: 10.3724/SP.J.1146.2013.01410
    [14] 宋相法, 焦李成.  基于稀疏编码和集成学习的多示例多标记图像分类方法, 电子与信息学报. 2013, 35(3): 622-626. doi: 10.3724/SP.J.1146.2012.01218
    [15] 李艳玲, 颜永红.  多特征融合的英语口语考试自动评分系统的研究, 电子与信息学报. 2012, 34(9): 2097-2102. doi: 10.3724/SP.J.1146.2012.00172
    [16] 施云飞, 宋千, 金添, 周智敏.  自适应特征筛选的地雷目标AdaBoost分类器, 电子与信息学报. 2011, 33(8): 1798-1802. doi: 10.3724/SP.J.1146.2010.01423
    [17] 潘泓, 李晓兵, 金立左, 夏良正.  一种基于二值粒子群优化和支持向量机的目标检测算法, 电子与信息学报. 2011, 33(1): 117-121. doi: 10.3724/SP.J.1146.2010.00260
    [18] 吴永辉, 计科峰, 李禹, 郁文贤.  利用SVM的极化SAR图像特征选择与分类, 电子与信息学报. 2008, 30(10): 2347-2351. doi: 10.3724/SP.J.1146.2007.00346
    [19] 方敏, 王宝树.  基于AdaBoost的改进模糊分类规则集成学习, 电子与信息学报. 2005, 27(5): 835-837.
    [20] 孙秀斌, 陈光禑, 谢永乐.  模拟集成电路的测试节点选择, 电子与信息学报. 2004, 26(4): 645-650.
  • 加载中
  • 图(11) / 表ll (4)
    计量
    • 文章访问数:  142
    • HTML全文浏览量:  89
    • PDF下载量:  10
    • 被引次数: 0
    出版历程
    • 收稿日期:  2020-01-09
    • 修回日期:  2020-06-15
    • 网络出版日期:  2020-07-22

    目录

      /

      返回文章
      返回

      官方微信,欢迎关注