首页 > 论文 > 光学学报 > 38卷 > 6期(pp:617001--1)

基于卷积神经网络与显微高光谱的胃癌组织分类方法研究

Gastric Carcinoma Classification Based on Convolutional Neural Network and Micro-Hyperspectral Imaging

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

为了探究高光谱技术在胃癌组织病理诊断中的应用,将高光谱成像与显微系统结合,采集胃部切片组织的高光谱图像。针对胃癌组织与胃部正常组织在410~910 nm波段的光谱特性差异,提出了一种基于卷积神经网络模型的胃癌组织分类方法,对原始光谱进行S-G平滑和一阶导数等预处理,通过分析光谱数据的特点和模型的分类效率,确定了最佳的网络结构及参数。实验结果表明:该模型对胃部癌变和正常组织的分类准确率为96.53%,鉴别胃癌组织的灵敏度和特异性分别达到94.29%和97.14%;相比于浅层学习方法,卷积神经网络模型能够充分提取癌变组织的深层光谱特征,同时能有效避免过拟合现象。将深度学习理论与显微高光谱结合的方法为医学病理研究提供了新思路。

Abstract

In order to explore the application of hyperspectral technology in the pathological diagnosis of gastric cancer, we combine hyperspectral imaging and microscopy to acquire hyperspectral images of gastric slices. According to the difference of spectral characteristics between gastric cancer tissue and normal gastric tissue in the wavelength of 410-910 nm, we propose a classification method based on convolutional neural network (CNN). The original spectrum is preprocessed by S-G smoothing and the first order derivative. We establish the optimal network structure and parameters by analyzing the spectral data characteristics and the classification efficiency. Experimental results show that the classification accuracy of cancerous and normal gastric tissues is 96.53%, the sensitivity and specificity of distinguishing gastric carcinoma reach 94.29% and 97.14%, respectively. Compared with shallow learning methods, the CNN model can fully extract the deep spectral characteristics of cancerous tissues and effectively prevent over-fitting. The method of deep learning combined with micro-hyperspectral imaging can also provide a new idea for the medical pathology research.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:O433.4

DOI:10.3788/aos201838.0617001

所属栏目:医用光学与生物技术

基金项目:国家重点研发计划(2017YFC1403700)、国家自然科学基金(61501456)、中国科学院光谱成像技术重点实验室开发基金(Y429J41213)

收稿日期:2017-12-13

修改稿日期:2018-01-19

网络出版日期:--

作者单位    点击查看

杜剑:中国科学院西安光学精密机械研究所光谱成像技术重点实验室, 陕西 西安 710119中国科学院大学, 北京 100049
胡炳樑:中国科学院西安光学精密机械研究所光谱成像技术重点实验室, 陕西 西安 710119
张周锋:中国科学院西安光学精密机械研究所光谱成像技术重点实验室, 陕西 西安 710119

联系人作者:胡炳樑(hbl@opt.ac.cn)

备注:杜剑(1991-),男,博士研究生,主要从事光谱数据处理方面的研究。E-mail: dujian@opt.cn

【1】Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012[J]. International Journal of Cancer, 2015, 136(5): 359-386.

【2】Vodinh T, Stokes D L, Wabuyele M B, et al. A hyperspectral imaging system for in vivo optical diagnostics[J]. IEEE Engineering in Medicine and Biology Magazine, 2004, 23(5): 40-49.

【3】Luo B, Zhang L. Robust autodial morphological profiles for the classification of high-resolution satellite images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2013, 52(2): 1451-1462.

【4】Hou B H, Yao M L, Wang R, et al. Spatial-spectral semi-supervised local discriminant analysis for heperspectral image classification[J]. Acta Optica Sinica, 2017, 37(7): 0728002.
侯榜焕, 姚敏立, 王榕, 等. 面向高光谱图像分类的空谱半监督局部判别分析[J]. 光学学报, 2017, 37(7): 0728002.

【5】Suzuki Y, Okamoto H, Takahashi M, et al. Mapping the spatial distribution of botanical composition and herbage mass in pastures using hyperspectral imaging[J]. Grassland Science, 2012, 58(1): 1-7.

【6】Hou H Y, Fang Z H, Zhang Y Z, et al. Simulation an in vivo experimental study on noninvasive spectral detection of skin cholesterol[J]. Chinese Journal of Lasers, 2016, 43(9): 0907001.
候华毅, 方朝晖, 张元志, 等. 皮肤胆固醇无创光谱检测模拟和在体实验研究[J]. 中国激光, 2016, 43(9): 0907001.

【7】Dong A G, Li J X, Zhang B, et al. Hyperspectral image classification algorithm based on spectral clustering and sparse representation[J]. Acta Optica Sinica, 2017, 37(8): 0828005.
董安国, 李佳逊, 张蓓, 等. 基于谱聚类和稀疏表示的高光谱图像分类算法[J]. 光学学报, 2017, 37(8): 0828005.

【8】Akbari H, Halig L V, Zhang H Z, et al. Detection of cancer metastasis using a novel macroscopic hyperspectral method[C]. SPIE, 2012, 8317: 831711.

【9】Liu H Y, Gu W Q, Li Q L, et al. Nerve classification with hyperspectral imaging technology[J]. Spectroscopy and Spectral Analysis, 2015, 35(1): 38-43.
刘洪英, 顾文荃, 李庆利, 等. 超光谱成像技术应用于神经分类的研究[J]. 光谱学与光谱分析, 2015, 35(1): 38-43.

【10】Zhu S, Su K, Liu Y, et al. Identification of cancerous gastric cells based on common features extracted from hyperspectral microscopic images[J]. Biomedical Optics Express, 2015, 6(4): 1135-1145.

【11】Akbari H, Halig L V, Schuster D M, et al. Hyperspectral imaging and quantitative analysis for prostate cancer detection[J]. Journal of Biomedical Optics, 2012, 17(7): 076005.

【12】Gerstner A O, Laffers W, Bootz F, et al. Hyperspectral imaging of mucosal surfaces in patients[J]. Journal of Biophotonics, 2012, 5(3): 255-262.

【13】Liu Z, Wang H, Li Q. Tongue tumor detection in medical hyperspectral images[J]. Sensors, 2012, 12(1): 162-174.

【14】Yuan J W, Zhang C G, Wang H, et al. Rapid microscopic spectral imaging of lung cancer tissue based on acousto-optic tunable filter[J]. Chinese Journal of Lasers, 2018, 45(4): 0407003.
原江伟, 张春光, 王号, 等. 基于声光可调滤波器的肺癌组织快速显微光谱成像[J]. 中国激光, 2018, 45(4): 0407003.

【15】Hinton G E, Salakhutdinov R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507.

【16】Lecun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[C]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.

【17】Karen S, Andrew Z. Very deep convolutional networks for large-scale image recognition[C]. Proceedings of ICLR, 2015: 1-14.

【18】Hinton G E, Srivastava N, Krizhevsky A, et al. Improving neural networks by preventing co-adaptation of feature detectors[J]. Computer Science, 2012, 3(4): 212-223.

【19】Zuo Z, Shuai B, Wang G, et al. Learning contextual dependence with convolutional hierarchical recurrent neural networks[J]. IEEE Transactions on Image Processing, 2016, 25(7): 2983-2996.

引用该论文

Du Jian,Hu Bingliang,Zhang Zhoufeng. Gastric Carcinoma Classification Based on Convolutional Neural Network and Micro-Hyperspectral Imaging[J]. Acta Optica Sinica, 2018, 38(6): 0617001

杜剑,胡炳樑,张周锋. 基于卷积神经网络与显微高光谱的胃癌组织分类方法研究[J]. 光学学报, 2018, 38(6): 0617001

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF