基于自监督学习的病理图像层次分割

doi:10.11772/j.issn.1001-9081.2019101863

计算机应用 ›› 2020, Vol. 40 ›› Issue (6): 1856-1862.DOI: 10.11772/j.issn.1001-9081.2019101863

• 应用前沿、交叉与综合 • 上一篇

基于自监督学习的病理图像层次分割

吴崇数^1,2, 林霖³, 薛蕴菁³, 时鹏^1,2

1.福建师范大学数学与信息学院，福州 350117
2.数字福建环境监测物联网实验室(福建师范大学)，福州 350117
3.福建医科大学附属协和医院放射科，福州 350001

收稿日期:2019-10-31 修回日期:2020-01-17 出版日期:2020-06-10 发布日期:2020-06-18
通讯作者: 时鹏(1980—)
作者简介:吴崇数(1996—)，男，浙江瑞安人，硕士研究生，主要研究方向：医学图像处理、机器学习.林霖(1988—)，男，福建泉州人，主治医师，博士研究生，主要研究方向：中枢神经放射学、影像组学、人工智能、分子影像.薛蕴菁(1971—)，女，浙江江阴人，教授，博士，主要研究方向：中枢神经影像、心血管影像、人工智能.时鹏(1980—)，男，河南许昌人，副教授，博士，CCF会员，主要研究方向：模式识别、医学图像处理、人工智能、机器学习
基金资助:
国家自然科学基金资助项目(61672157)；福建省科技创新联合资金资助项目(2018Y9112，2018Y9044)；福建省卫生健康中青年骨干人才培养项目(2019-ZQN-17)。

Hierarchical segmentation of pathological images based on self-supervised learning

WU Chongshu^1,2, LIN Lin³, XUE Yunjing³, SHI Peng^1,2

1. College of Mathematics and Informatics, Fujian Normal University, Fuzhou Fujian 350117, China
2. Digit Fujian Internet-of-Things Laboratory of Environmental Monitoring (Fujian Normal University), Fuzhou Fujian 350117, China
3. Radiology Department, Fujian Medical University Union Hospital, Fuzhou Fujian 350001, China

Received:2019-10-31 Revised:2020-01-17 Online:2020-06-10 Published:2020-06-18
Contact: SHI Peng, born in 1980, Ph. D., associate professor. His research interests include pattern recognition, medical image processing, artificial intelligence, machine learning.
About author:WU Chongshu, born in 1996, M. S. candidate. His research interests include medical image processing, machine learning.LIN Lin, born in 1988, Ph. D. candidate, chief physician. His research interests include central neuroradiology, radiomics， artificial intelligence, molecular imaging.XUE Yunjing, born in 1971, Ph. D., professor. Her research interests include central nervous imaging, cardiovascular imaging, artificial intelligence.SHI Peng, born in 1980, Ph. D., associate professor. His research interests include pattern recognition, medical image processing, artificial intelligence, machine learning.
Supported by:
National Natural Science Foundation of China (61672157), the Science and Technology Innovation Joint Foundation of Fujian Province (2018Y9112，2018Y9044), the Development Program of Young and Middle-aged Talents of Health of Fujian Province (2019-ZQN-17).

摘要/Abstract

摘要： 在苏木精-伊红(HE)染色病理图像中，细胞染色分布的不均匀和各类组织形态的多样性给病理图像的自动分割带来极大挑战。为解决该问题，提出了一种基于自监督学习的病理图像三步层次分割方法，对病理图像中各类组织进行由粗略到精细的全自动逐层分割。首先，根据互信息的计算结果在RGB色彩空间中进行特征选择；其次，采用K-means聚类将图像初步分割为各类组织结构的色彩稳定区域与模糊区域；然后，以色彩稳定区域为训练集采用朴素贝叶斯分类对模糊区域进行进一步分割，得到完整的细胞核、细胞质和胞外间隙这三类组织结构；最后，对细胞核部分进行结合形状和色彩强度的混合分水岭分割得到细胞核间的精确边界，进而量化计算细胞核个数、核占比、核质比等指标。对脑膜瘤HE染色病理图像的分割实验结果表明，所提方法对于染色和细胞形态差异保持较高的鲁棒性，各类组织区域分割误差在5%以内，在细胞核分割精度的对比实验中平均正确率在96%以上，满足临床自动图像分析的要求，其量化结果可以为定量病理分析提供依据。

关键词: 病理图像, 图像分割, 自监督学习, K-means聚类, 朴素贝叶斯分类

Abstract: The uneven distribution of cell staining and the diversity of tissue morphologies bring challenges to the automatic segmentation of Hematoxylin-Eosin (HE) stained pathological images. In order to solve the problem, a three-step hierarchical segmentation method of pathological images based on self-supervised learning was proposed to automatically segment the tissues in the pathological images layer-by-layer from coarse to fine. Firstly, feature selection was performed in the RGB color space based on the calculation result of mutual information. Secondly, the image was initially segmented into stable and fuzzy color regions of each tissue structure based on K-means clustering. Thirdly, the stable color regions were taken as training datasets for further segmentation of fuzzy color regions by naive Bayesian classification, and the three complete tissue structures including nucleus, cytoplasm and extracellular space were obtained. Finally, precise boundaries between nuclei were obtained by performing the mixed watershed classification considering both shape and color intensities to the nucleus part, so as to quantitatively calculate the indicators such as the number of nuclei, nucleus proportion, and nucleus-cytoplasm ratio. Experimental results of HE stained meningioma pathological image segmentation show that， the proposed method is highly robust to the difference of staining and cell morphologies, the error of issue area segmentation is within 5%, and the average accuracy of the proposed method in nucleus segmentation accuracy experiment is above 96%, which means that the proposed method can meet the requirements of automatic analysis of clinical images and its quantitative results can provide references for quantitative pathological analysis.

Key words: pathological image, image segmentation, self-supervised learning, K-means clustering, naive Bayesian classification

中图分类号:

TP181

吴崇数, 林霖, 薛蕴菁, 时鹏. 基于自监督学习的病理图像层次分割[J]. 计算机应用, 2020, 40(6): 1856-1862.

WU Chongshu, LIN Lin, XUE Yunjing, SHI Peng. Hierarchical segmentation of pathological images based on self-supervised learning[J]. Journal of Computer Applications, 2020, 40(6): 1856-1862.

参考文献

1 TSOIK K F, CHANF C H, HIRAIH W, et al. Data visualization with IBM Watson analytics for global cancer trends comparison from world health organization [J]. International Journal of Healthcare Information Systems and Informatics, 2018, 13(1): 45-54.
2 GURCANM N, BOUCHERONL E, CAN A, et al. Histopathological image analysis: a review [J]. IEEE Reviews in Biomedical Engineering, 2009, 2: 147-171.
3 BELSAREA D, MUSHRIFM M. Histopathological image analysis using image processing techniques: an overview [J]. Signal and Image Processing: An International Journal, 2012, 3(4):23-36.
4 ÇAYIRS, SERTELIE A, AYALTIS, et al. Segmentation of the main structures in hematoxylin and eosin images [C]// Proceedings of the 26th Signal Processing and Communications Applications Conference. Piscataway: IEEE, 2018: 1-4.
5 BELSAREA D, MUSHRIFM M, PANGARKARM A, et al. Breast histopathology image segmentation using spatio-colour-texture based graph partition method [J]. Journal of Microscopy, 2016, 262(3):260-273.
6 JOTHIJ A A, RAJAMV M A. Segmentation of nuclei from breast histopathology images using PSO-based Otsu’s multilevel thresholding [M]// SURESH L P, DASH S S, PANIGRAHI B K. Artificial Intelligence and Evolutionary Algorithms in Engineering Systems, AISC 325. New Deli: Springer, 2015: 835-843.
7 STYNERM A, ANGELINIE D, TAHERIS, et al. Robust nuclei segmentation in cyto-histopathological images using statistical level set approach with topology preserving constraint [C]// Proceedings of the SPIE 10133, Medical Imaging 2017: Image Processing. Bellingham, WA: SPIE, 2017: Article No.1013318.
8 BEEVIK S, NAIRM S, BINDUG R. Automatic segmentation of cell nuclei using krill herd optimization based multi-thresholding and localized active contour model [J]. Biocybernetics and Biomedical Engineering, 2016, 36(4): 584-596.
9 WIENERTS, HEIMD, SAEGERK, et al. Detection and segmentation of cell nuclei in virtual microscopy images: a minimum-model approach [J]. Scientific Reports, 2012, 2: Article No.503.
10 SONGY, ZHANGL, CHENS, et al. Accurate segmentation of cervical cytoplasm and nuclei based on multiscale convolutional network and graph partitioning [J]. IEEE Transactions on Biomedical Engineering, 2015, 62(10): 2421-2433.
11 ISAKSSONJ, ARVIDSSONI, ÅASTRÖMK, et al. Semantic segmentation of microscopic images of H&E stained prostatic tissue using CNN [C]// Proceedings of the 2017 IEEE International Joint Conference on Neural Networks. Piscataway: IEEE, 2017: 1252-1256.
12 CHENH, QIX, YUL, et al. DCAN: deep contour-aware networks for object instance segmentation from histology images [J]. Medical Image Analysis, 2017, 36: 135-146.
13 SUH, LIUF, XIEY, et al. Region segmentation in histopathological breast cancer images using deep convolutional neural network [C]// Proceedings of the IEEE 12th International Symposium on Biomedical Imaging. Piscataway: IEEE, 2015: 55-58.
14 SUROVA, GINATD T, LIM T, et al. Histogram analysis parameters apparent diffusion coefficient for distinguishing high and low-grade meningiomas: a multicenter study [J]. Translational Oncology, 2018, 11(5): 1074-1079.
15 KIERNANJ A. Histological and Histochemical Methods [M]. 4th ed. Banbury: Scion Publishing, 2008: 481-495.
16 SUNT, NEUVOY. Detail-preserving median based filters in image processing [J]. Pattern Recognition Letters, 1994, 15(4): 341-347.
17 VLIET L JVAN, YOUNGI T, VERBEEKP W. Recursive Gaussian derivative filters [C]// Proceedings of the 14th International Conference on Pattern Recognition. Piscataway: IEEE, 1998:509-514.
18 GAOW, HUL, ZHANGP. Class-specific mutual information variation for feature selection [J]. Pattern Recognition, 2018, 79: 328-339.
19 SHIP, ZHONGJ, HUANGR, et al. Automated quantitative image analysis of hematoxylin-eosin staining slides in lymphoma based on hierarchical Kmeans clustering [C]// Proceedings of the 8th International Conference on Information Technology in Medicine and Education. Piscataway: IEEE, 2017: 99-104.
20 SHIP, CHENJ, LINJ, et al. High-throughput fat quantifications of hematoxylin-eosin stained liver histopathological images based on pixel-wise clustering [J]. SCIENCE CHINA Information Sciences, 2017, 60(9): Article No.092108.
21 BHAGYA SHREES R, SHESHADRIH S. Diagnosis of Alzheimer’s disease using naive Bayesian classifier [J]. Neural Computing and Applications, 2018, 29(1): 123-132.
22 KARVELISP S, TZALLASA T, FOTIADISD I, et al. A multichannel watershed-based segmentation method for multispectral chromosome classification [J]. IEEE Transactions on Medical Imaging, 2008, 27(5): 697-708.
23 VETAM, HUISMANA, VIERGEVERM A, et al. Marker-controlled watershed segmentation of nuclei in H&E stained breast cancer biopsy images [C]// Proceedings of the 2011 IEEE International Symposium on Biomedical Imaging: from Nano to Macro. Piscataway: IEEE, 2011: 618-621.
24 COLLINST. ImageJ for medical microscopy image processing: an introduction to macro development for batch processing [M]// CRISTOBAL G, SCHELKENS P, THIENPONT H. Optical and Digital Image Processing: Fundamentals and Applications. Rastede: Wiley-VCH, 2011: 859-877.
25 HEDBERG-BUENZA, CHRISTOPHERM A, LEWISC J, et al. RetFM-J, an ImageJ-based module for automated counting and quantifying features of nuclei in retinal whole-mounts [J]. Experimental Eye Research, 2015, 146: 386-392.
26 CHATTARAJA, DAS A, BHATTACHARYAM. Mammographic image segmentation by marker controlled watershed algorithm [C]// Proceedings of the 2017 IEEE International Conference on Bioinformatics and Biomedicine. Piscataway: IEEE, 2017: 1000-1003.
27 HUOC, HONGH, ZHANGL, et al. Adhesion cells segmentation based on distance map and watershed transform [J]. ICIC Express Letters, Part B, Applications: an International Journal of Research and Surveys, 2015, 6(12): 3243-3249.

[1]	代雨柔, 杨庆, 张凤荔, 周帆. 基于自监督学习的社交网络用户轨迹预测模型[J]. 计算机应用, 2021, 41(9): 2545-2551.
[2]	杨瑞, 钱晓军, 孙振强, 许振. 自然场景下多区域特征融合的混合航拍图像分割算法[J]. 计算机应用, 2021, 41(8): 2445-2452.
[3]	沈雪雯, 王晓东, 姚宇. 基于空间分频的超声图像分割注意力网络[J]. 计算机应用, 2021, 41(6): 1828-1835.
[4]	黄梨, 卢龙. 基于长距离依赖编码与深度残差U-Net的缺血性卒中病灶分割[J]. 计算机应用, 2021, 41(6): 1820-1827.
[5]	罗琴, 王艳. 无需初始轮廓的图像分割模型[J]. 计算机应用, 2021, 41(4): 1179-1183.
[6]	董阳, 潘海为, 崔倩娜, 边晓菲, 滕腾, 王邦菊. 面向多模态磁共振脑瘤图像的小样本分割方法[J]. 计算机应用, 2021, 41(4): 1049-1054.
[7]	胡屹杉, 秦品乐, 曾建潮, 柴锐, 王丽芳. 基于特征融合和动态多尺度空洞卷积的超声甲状腺分割网络[J]. 计算机应用, 2021, 41(3): 891-897.
[8]	袁芊芊, 邓洪敏, 王晓航. 基于超像素快速模糊C均值聚类与支持向量机的柑橘病虫害区域分割[J]. 计算机应用, 2021, 41(2): 563-570.
[9]	孙中杰, 万涛, 陈东, 汪昊, 赵艳丽, 秦曾昌. 深度学习在主动脉中膜变性病理图像分类中的应用[J]. 计算机应用, 2021, 41(1): 280-285.
[10]	原泉, 王艳, 李玉先. 距离保持水平集演化模型的快速实现算法[J]. 计算机应用, 2020, 40(9): 2743-2747.
[11]	高海军, 曾祥银, 潘大志, 郑伯川. 基于U-Net改进模型的直肠肿瘤分割方法[J]. 计算机应用, 2020, 40(8): 2392-2397.
[12]	曾招鑫, 刘俊. 基于深度学习的秀丽隐杆线虫显微图像分割方法[J]. 计算机应用, 2020, 40(5): 1453-1459.
[13]	张一飞, 李新福, 田学东. 基于图像分割的立体匹配算法[J]. 计算机应用, 2020, 40(5): 1415-1420.
[14]	魏小娜, 邢嘉祺, 王振宇, 王颖珊, 石洁, 赵地, 汪红志. 基于改进U-Net的关节滑膜磁共振图像的分割[J]. 计算机应用, 2020, 40(11): 3340-3345.
[15]	冯涛, 陈斌, 张跃飞. 基于改进的Mask R-CNN的染色体图像分割框架[J]. 计算机应用, 2020, 40(11): 3332-3339.

基于自监督学习的病理图像层次分割

Hierarchical segmentation of pathological images based on self-supervised learning

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics