Content-based image retrieval applied to BI-RADS tissue classification in screening mammography  

作　　者：Júlia Epischina Engrácia de Oliveira Arnaldo de Albuquerque Araújo Thomas M Deserno 

机构地区：[1]Department of Computer Science,Universidade Federal de Minas Gerais,31270-901,Belo Horizonte,MG,Brazil [2]Department of Medical Informatics,RWTH Aachen University,52074,Aachen,Germany

出　　处：《World Journal of Radiology》2011年第1期24-31,共8页世界放射学杂志（英文版）（电子版）

基　　金：Supported by CNPq-Brazil,Grants 306193/2007-8,471518/ 2007-7,307373/2006-1 and 484893/2007-6,by FAPEMIG,Grant PPM 347/08,and by CAPES;The IRMA project is funded by the German Research Foundation(DFG),Le 1108/4 and Le 1108/9

摘　　要：AIM:To present a content-based image retrieval(CBIR) system that supports the classification of breast tissue density and can be used in the processing chain to adapt parameters for lesion segmentation and classification.METHODS:Breast density is characterized by image texture using singular value decomposition(SVD) and histograms.Pattern similarity is computed by a support vector machine(SVM) to separate the four BI-RADS tissue categories.The crucial number of remaining singular values is varied(SVD),and linear,radial,and polynomial kernels are investigated(SVM).The system is supported by a large reference database for training and evaluation.Experiments are based on 5-fold cross validation.RESULTS:Adopted from DDSM,MIAS,LLNL,and RWTH datasets,the reference database is composed of over 10000 various mammograms with unified and reliable ground truth.An average precision of 82.14% is obtained using 25 singular values(SVD),polynomial kernel and the one-against-one(SVM).CONCLUSION:Breast density characterization using SVD allied with SVM for image retrieval enable the development of a CBIR system that can effectively aid radiologists in their diagnosis.AIM: To present a content-based image retrieval (CBIR) system that supports the classification of breast tissue density and can be used in the processing chain to adapt parameters for lesion segmentation and classification.METHODS: Breast density is characterized by image texture using singular value decomposition (SVD) and histograms. Pattern similarity is computed by a support vector machine (SVM) to separate the four BI-RADS tissue categories. The crucial number of remaining singular values is varied (SVD), and linear, radial, and polynomial kernels are investigated (SVM). The system is supported by a large reference database for training and evaluation. Experiments are based on 5-fold cross validation.RESULTS: Adopted from DDSM, MIAS, LLNL, and RWTH datasets, the reference database is composed of over 10 000 various mammograms with unified and reliable ground truth. An average precision of 82.14% is obtained using 25 singular values (SVD), polynomial kernel and the one-against-one (SVM).CONCLUSION: Breast density characterization using SVD allied with SVM for image retrieval enable the development of a CBIR system that can effectively aid radiologists in their diagnosis.

关 键 词：COMPUTER-AIDED diagnosis CONTENT-BASED IMAGE retrieval IMAGE processing Screening MAMMOGRAPHY SINGULAR value decomposition Support vector machine 

分 类 号：R655.8[医药卫生—外科学]