Discrimination between benign and malignant breast cancers in ultrasound images based on cost-sensitive boosting

Authors:
Xing Shen;Shuheng Zhang;Rui Yao;Yaqing Chen;Yue-Min Zhu;Su Zhang
Affiliations:
School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China;School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China;School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China;Department of Ultrasound, Xinhua Hospital, Shanghai, China;CREATIS, CNRS UMR5220, Inserm U1044, INSA Lyon, Villeurbanne, France;School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China
Venue:
IScIDE'11 Proceedings of the Second Sino-foreign-interchange conference on Intelligent Science and Intelligent Data Engineering
Year:
2011

Citing 6
Cited 0

Filters, Wrappers and a Boosting-Based Hybrid for Feature Selection

ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning
Minimum Redundancy Feature Selection from Microarray Gene Expression Data

CSB '03 Proceedings of the IEEE Computer Society Conference on Bioinformatics
Shape symmetry analysis of breast tumors on ultrasound images

Computers in Biology and Medicine
Feature selection using fuzzy entropy measures with similarity classifier

Expert Systems with Applications: An International Journal
Cost-Sensitive Boosting

IEEE Transactions on Pattern Analysis and Machine Intelligence
Adapt the mRMR criterion for unsupervised feature selection

ADMA'10 Proceedings of the 6th international conference on Advanced data mining and applications - Volume Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

Breast cancer is one of the high-risk cancers, and breast ultrasound is routinely used as an adjunct to mammography for detection and diagnosis. Furthermore, the effective computer-aided diagnosis (CAD) system could improve the specificity of discriminating malignant from benign lesions on breast ultrasound images. This paper presents a method for discrimination between benign and malignant breast cancers in ultrasound images based on cost-sensitive boosting. Firstly, the image feature is extracted according to BI-RADS (Breast imaging report and data system), and a more simplified sub-feature set is obtained through minimal redundancy maximal relevance (mRMR) algorithm. Then three cost-sensitive Boosting models are trained and compared, and the optimal classification parameters are obtained by cross validation. Experiment shows that cost-sensitive AdaBoost performs the best, with AUC (area under receive operating characteristic curve) at 0.859 in the condition of controlled FNR (false negative rate) at 5%, better than CS-RealBoost and CS-LogitBoost.