基于CT图像的肾脏肿瘤纹理特征提取
|
摘要
目前计算机辅助检测肿瘤都是基于病变形态学变化的分析,且这些算法的效果难以满足现状。从这些算法所忽略的图像纹理特征出发,从不同病人的198张、5类不同的良恶性肿瘤的CT图像中,基于灰度共生矩阵和灰度梯度共生矩阵,综合考虑肾脏肿瘤没有明显方向性及细纹理的特性,并依据可区分性、唯一性、不相关性以及为避免后续肿瘤识别过程复杂化,通过分析作出了有效性选择,首次提取出最能体现5种肿瘤的27个特征并验证其有效性,作为后续计算机辅助识别肾脏肿瘤研究的基础。
At present, computer-aided detection of tumors is based on the analysis of pathological changes, and the effect of these algorithms is unsatisfactory. Starting from the image texture features neglected by these algorithms, based on the gray level co-occurrence matrix and the gray level gradient co-occurrence matrix, from 198 CT images of different patients of 5 different types of benign and malignant tumors, the characteristics of renal tumors with no obvious orientation and fine texture are considered comprehensively, and according to distinguishability, uniqueness and irrelevance, the complexity of the subsequent tumor recognition process is avoided. Through the analysis, the effectiveness selection is made. First, 27 features of five types of tumors are extracted, and their effectiveness is verified as basis for the research on the follow-up computer-aided recognition of kidney tumors.
At present, computer-aided detection of tumors is based on the analysis of pathological changes, and the effect of these algorithms is unsatisfactory. Starting from the image texture features neglected by these algorithms, based on the gray level co-occurrence matrix and the gray level gradient co-occurrence matrix, from 198 CT images of different patients of 5 different types of benign and malignant tumors, the characteristics of renal tumors with no obvious orientation and fine texture are considered comprehensively, and according to distinguishability, uniqueness and irrelevance, the complexity of the subsequent tumor recognition process is avoided. Through the analysis, the effectiveness selection is made. First, 27 features of five types of tumors are extracted, and their effectiveness is verified as basis for the research on the follow-up computer-aided recognition of kidney tumors.
引文
[1]Leef JL,Klein JS.The Solitary Pulmonary Nodule[J].Radiol Clin North Am,2002,40(1):123-143.
[2]Sortini D,Maravegias K,Sortini A.Difficulty of Early Diagnosis in Patients with Solitary Pulmonary Nodule[J].Thorac Cardiovasc Surg,2005,129:1196.
[3]Sertel O,Lozanski G,Shana'ah A,et al.Computer-aided detection of centroblasts for follicular lymphoma grading using adaptive likelihood-based cell segmentation[J].IEEETransactions on Biomedical Engineering,2010,57(10):2613-2616.
[4]Mikhaylov VV,Bakhshiev AV.The System for Histopathology Images Analysis of Spinal Cord Slices[J].Procedia Computer Science,2017,103:239-243.
[5]Nabizadeh N,Kubat M.Brain tumors detection and segmentation in MR images:Gabor wavelet vs.statistical features[J].Computers&Electrical Engineering,2015(45):286-301.
[6]Beyer MH.The GLCM Tutorial[EB/OL].http://www.fp.ucalgary.ca/mhallbey/tutorial.htm,Accessed on 29 March,2017.
[7]伯特霍尔德·霍恩.机器视觉[M].蒋欣兰,译.北京:中国青年出版社,2014.
[8]Haralick RM,Shanmugam K,Dinstein I.Textural features for image classification[J].IEEE Transactions on Systems,Man,and Cybernetics,1973(SMC-3):610-621.
[9]Parvez A,Phadke AC.Effcient Implementation of GLCM based Texture Feature Computation using CUDA Platform[C]//International Conference on Trends in Electronics and Informatics,2017:296-300.
[10]http://www.ilovematlab.cn/thread-64831-1-1.html.
[2]Sortini D,Maravegias K,Sortini A.Difficulty of Early Diagnosis in Patients with Solitary Pulmonary Nodule[J].Thorac Cardiovasc Surg,2005,129:1196.
[3]Sertel O,Lozanski G,Shana'ah A,et al.Computer-aided detection of centroblasts for follicular lymphoma grading using adaptive likelihood-based cell segmentation[J].IEEETransactions on Biomedical Engineering,2010,57(10):2613-2616.
[4]Mikhaylov VV,Bakhshiev AV.The System for Histopathology Images Analysis of Spinal Cord Slices[J].Procedia Computer Science,2017,103:239-243.
[5]Nabizadeh N,Kubat M.Brain tumors detection and segmentation in MR images:Gabor wavelet vs.statistical features[J].Computers&Electrical Engineering,2015(45):286-301.
[6]Beyer MH.The GLCM Tutorial[EB/OL].http://www.fp.ucalgary.ca/mhallbey/tutorial.htm,Accessed on 29 March,2017.
[7]伯特霍尔德·霍恩.机器视觉[M].蒋欣兰,译.北京:中国青年出版社,2014.
[8]Haralick RM,Shanmugam K,Dinstein I.Textural features for image classification[J].IEEE Transactions on Systems,Man,and Cybernetics,1973(SMC-3):610-621.
[9]Parvez A,Phadke AC.Effcient Implementation of GLCM based Texture Feature Computation using CUDA Platform[C]//International Conference on Trends in Electronics and Informatics,2017:296-300.
[10]http://www.ilovematlab.cn/thread-64831-1-1.html.