基于深度学习算法的脑肿瘤CT图像特征分割技术改进Improvement of brain tumor CT image feature segmentation technology based on in-depth learning algorithm
崔仲远,黄伟
摘要(Abstract)：
针对基于数学形态的脑肿瘤CT图像特征分割技术存在准确率低、分割效果不明确的弊端,提出基于深度学习算法的脑肿瘤CT图像特征分割技术。将可视人体数据集CVH-2作为研究对象,对数据集中的图像实施预处理,对图像四个模态实施卷积分别获取不同模态彼此的差异信息,归一化获取脑肿瘤CT图像多模态3D-CNNs特征。对基于SAE深度学习算法的脑肿瘤CT图像特征分割模型实施二级训练,将脑肿瘤CT图像多模态3D-CNNs特征经过处理后获取的S,V通道数据输入模型实施训练,在第二级训练的过程中把第一级SAE训练得到的权重作为二级训练的原始权重,将一级训练中错误分割的组织结构和沟回作为二次训练的数据集,获取脑肿瘤CT图像特征的准确分割结果。实验结果表明,所提方法在脑肿瘤CT图像特征分割准确率和效率方面具有显著优势。
关键词(KeyWords)： 深度学习算法;脑肿瘤CT图像;特征分割技术;多模态3D-CNN;SAE结构;数据集
基金项目(Foundation): 国家自然科学基金(61602423);; 2018年度河南省科技厅科技攻关项目:基于深度学习的脑肿瘤图像的分割方法(182102310694)~~
作者(Author): 崔仲远,黄伟
DOI: 10.16652/j.issn.1004-373x.2018.16.023
参考文献(References)：
- [1]马成虎,董洪伟.一种基于深度学习的多尺度深度网络的场景标注算法[J].计算机工程与科学,2016,38(7):1356-1361.MA Chenghu,DONG Hongwei.A scene labeling algorithm for multi-scale deep networks based on deep learning[J].Computer engineering and science,2016,38(7):1356-1361.
- [2]徐立,白金牛,李磊民.基于脑部MR图像GMM特征决策分类的肿瘤诊断[J].控制工程,2017,24(8):1718-1722.XU Li,BAI Jinniu,LI Leimin.Diagnosis of tumor in brain MR images based on GMM features and decision tree classifier[J].Control engineering of China,2017,24(8):1718-1722.
- [3]ROOBINI K,RAMAKRISHNAN T,SANKARAGOMATHI B.Neighbourhood weighted fuzzy C-means(NWFCM)clustering algorithm for segmentation of brain tumors in CT images[J].International journal of applied engineering research,2015,10(20):15327-15330.
- [4]秦丰,刘东霞,孙炳达,等.基于深度学习和支持向量机的4种苜蓿叶部病害图像识别[J].中国农业大学学报,2017,22(7):123-133.QIN Feng,LIU Dongxia,SUN Bingda,et al.Image recognition of four different alfalfa leaf diseases based on deep learning and support vector machine[J].Journal of China Agricultural University,2017,22(7):123-133.
- [5]张人上.基于NSCT-GLCM的CT图像特征提取算法[J].计算机工程与应用,2014,50(11):159-162.ZHANG Renshang.Features extraction algorithm of CT image based on GNSCT-LCM[J].Computer engineering and applications,2014,50(11):159-162.
- [6]楚陪陪,魏本征,曲彦,等.基于复合超像素技术的肺部CT图像分割算法[J].合肥工业大学学报(自然科学版),2017,40(3):332-339.CHU Peipei,WEI Benzheng,QU Yan,et al.A lung CT image segmentation algorithm based on composite super pixel technology[J].Journal of Hefei University of Technology(Natural science),2017,40(3):332-339.
- [7]汪亮,金福江,陈峻严.基于区域生长和FCM模糊聚类的颅内出血CT图像分割[J].系统仿真学报,2014,26(2):231-235.WANG Liang,JIN Fujiang,CHEN Junyan.Segmentation of intracranial hemorrhage′s CT image based on method of region growing and FCM fuzzy clustering[J].Journal of system simulation,2014,26(2):231-235.
- [8]CHITRA K,RAMAKRISHNAN T,SANKARAGOMATHI B.Automatic segmentation of brain tumor in CT images using LIPC[J].International journal of applied engineering research,2015,10(20):15478-15483.
- [9]杨阳,张文生.基于深度学习的图像自动标注算法[J].数据采集与处理,2015,30(1):88-98.YANG Yang,ZHANG Wensheng.Image auto-annotation based on deep learning[J].Journal of data acquisition&processing,2015,30(1):88-98.
- [10]王博,郭继昌,张艳.基于深度网络的可学习感受野算法在图像分类中的应用[J].控制理论与应用,2015,32(8):1114-1119.WANG Bo,GUO Jichang,ZHANG Yan.Learnable receptive fields scheme in deep networks for image categorization[J].Control theory&applications,2015,32(8):1114-1119.