北京生物医学工程

卷积神经网络在肝癌病理切片图像分类中的应用

Application of convolutional neural network in image classification of liver cancer pathological section

作者：茹仙古丽·艾尔西丁艾尔潘江·库德来提严传波姚娟

单位：1新疆医科大学基础医学院（乌鲁木齐830011） 2 新疆医科大学医学工程技术学院（乌鲁木齐830011） 3 新疆医科大学第一附属医院放射科（乌鲁木齐830054）

关键词：卷积神经网络；肝脏组织切片图像； Inception V3；图像分类

分类号：R318.04；TP751

出版年·卷·期（页码）：2020·39·1（29-33）

摘要：

目的探讨基于卷积神经网络的肝脏组织切片图像正常和病变性分类方法的可行性及应用价值。方法使用一种能够自动学习图像特征并分类的方法，先利用原始的Inception V3 模型对肝脏组织切片数据集上进行训练，然后在原始模型的基础上通过微调得到改进的Inception V3模型，最后用改进的模型来实现肝脏组织切片图像正常和病变性两种类型的分类。结果改进后的Inception V3模型对肝脏切片图像的分类结果较佳，平均分类准确率达到99.2%。结论卷积神经网络的肝脏组织切片图像正常和病变性分类方法可行、合理，改进的Inception V3模型的分类效果较好。

Objective To investigate the feasibility and application value of normal and pathological classification of liver tissue slices based on convolution neural network.Methods Using a method that can automatically learn image features and classify, we first use the original inception v3 model to train the liver tissue slice data set, then we can get the improved inception v3 model by fine-tuning the original model. At last, we use the improved model to realize the classification of liver tissue slice image of normal and pathological two types.Results The improved Inception V3 model has better classification results for liver slice images with an average Average classification accuracy of 99.2%.Conclusions The normal and pathological classification of the liver tissue slice images of the convolution neural network is feasible and reasonable, and the improved classification effect of the Inception V3 model is suitable.

参考文献：

[1]Chen WQ,Zheng RS,Baade, Peter D,et a1.Cancer statistics in China,2015[J].CA: A Cancer Journal For Clinicians,2016,66(2):115-132.

[2]Khazaei Z,Sohrabivafa M,Momenaba V.Global cancer statistics 2018: Globocan estimates of incidence and mortality worldwide prostate cancers and their relationship with the human development index[J].Advances in Human Biology,2019,9(3):245-250.

[3]杨秉辉，任正刚，汤钊猷.关于肝癌诊断与分期标准的讨论[J].中华肝癌病杂志，2000,8（3）：133-134.

[4]Kuppili V,Biswas M,Sreekumar A,et al.Extreme learning machine framework for risk stratification of fatty liver disease using ultrasound tissue characterization[J].Journal Of Medical Systems,2017,41(10):152.

[5] Conze PH,Noblet V,Rousseau F,et al.Scale-adaptive supervoxel-based random forests for liver tumor segmentation in dynamic contrast-enhanced CT scans[J].International Journal Of Computer Assisted Radiology And Surgery,2017,12(2):223-233.

[6]郝涛，张智.基于BP神经网络的原发性肝癌CT图像纹理分析[J].中国数字医学，2013，8(8)：73-76.

Hao T,Zhang Z.Primary hepatic carcinoma texture analysis based on CT image[J].China Digital Medicine,2013，8(8)：73-76

[7]刘建华，王建伟.基于图像处理的CT图像肝癌诊断技术研究[J].清华大学学报，2014，54(7)：917-923.

Liu JH,Wang JW.Liver cancer diagnosis based on CT image processing[J].Journal Of Qinghua University,2014,54(7)：917-923.

[8]祁亮，沈洁.机器学习在肝癌诊疗领域的应用进展[J].癌症进展，2019,17（5）：519-525.

Yan L, Shen J.Progress in the application of machine learning in the field of liver cancer diagnosis and treatment[J].Oncology Progress,2019,17（5）:519-525.

[9]Rajpurkar P, Hannun AY, Haghpanahi M, et al. Cardiologist-level arrhythmia detection with convolutional neural networks[EB/OL].[2018-10-23]. http://cn.arxiv.org/abs/1707.01836.

[10]Arevalo J, Gonzalez FA, Ramos-Pollan R, et al. Convolutional neural networks for mammography mass lesion classification [C]//37th Annual International Conference Of The IEEE Engineering In Medicine & Biology Society. Milan: IEEE, 2015: 797.

[11]赵鹏飞，赵涓涓，强彦，等.多输入卷积神经网络肺结节检测方法研究[J].计算机科学，2018,45（1）：162-166.

Zhao PF,Zhao JJ,Qiang Y,et al.Study on detection method of pulmonary nodules with multiple input convolution neural network[J].Computer Science,2018,45(1):162-166.

[12]王岩.卷积神经网络在肝穿刺图像分类中的应用[J].电脑知识与技术，2018, 14(25)：203-205.

Wang Y.Application of onvolution neural network in image classification of liver puncture[J].Computer Knowledge And Technology,2018, 14(25)：203-205.

[13]刘飞，张俊然，杨豪.基于深度学习的医学图像识别研究进展[J].中国生物医学工程学报，2018, 37(1):86-94.
Liu F,Zhang JR,Yang H.Research progress of medical image recognition based on deep learning[J].Chinese Journal Of Biomedical Engineering,2018, 37(1):86-94.

[14]张晴晴, 刘勇, 王智超,等.卷积神经网络在语音识别中的应用[J].网络新媒体技术, 2014, 3(6)：39-42.

Zhang QQ, Liu Y, Wang ZC.Application of convolutional neural network in speech recognition[J].Network New Media Technology,2014, 3(6)：39-42.

[15]张文宇，刘畅.卷积神经网络算法在语音识别中的应用[J].信息技术，2018, (10)：147-152.

Zhang WY,Liu C.Application of convolutional neural network algorithm in speech recognition[J].Information Technology,2018, (10)：147-152.

[16]赵显达，黄欢.基于卷积神经网络的人脸识别的研究[J].信息技术，2018, (9)：15-19,23.

Zhao XD,Huang H.Research on face recognition based on convolutional neural network[J].Information Technology,2018, (9)：15-19,23.

[17]陆红.基于卷积神经网络人脸识别方法研究[J].现代信息科技，2018, 2(10)：102-103,106.

Lu H.Face Recognition method base on convolution neural network[J].Modern Informationn Technology,2018, 2(10)：102-103,106.

[18]丁冠祺，刘宇涵，杨皓博.基于卷积神经网络的人脸识别[J].信息记录材料，2018, 19(9)：48-49.

Ding GQ,Liu YH,Yang HB.Face recognition based on convolutional neural network[J].Information Recording Materials,2018, 19(9)：48-49.

[19]Park E, Han X, Berg T L, et al. Combining multiple sources of knowledge in deep CNNs for action recognition[C]// IEEE Winter Conference On Applications of Computer Vision. IEEE Computer Society, 2016:1-8.

[20]蔡国永, 夏彬彬.基于卷积神经网络的图文融合媒体情感预测[J].计算机应用, 2016, 36(2)：428-431.

Cai GY,Xia BB.Multimedia sentiment analysis based on convolutional neural network[J].Journal Of Computer Applications,2016, 36(2)：428-431.

[21]Yosinski J,Clune J，Benjio Y,et a1．How transferable are features in deep neural networks? [J].EprintArxiv,2014,27: 3320-3328．

[22]邹铁.基于深度卷积网络的图像分类算法研究[J].安徽电子信息职业技术学院学报,2017，16（6）：203-208.

Zou T.On the image classification algorithm based on deep convolution neural network[J].Journal Of Anhui Vocational College Of Electronics And Information Technology,2017,16（6）:203-208.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录