2017 IEEE International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET 2017)

190 / 2016-12-29 02:58:20

Breast cancer detection using thermography

Breast cancer ,1103,Thermography,Classifier

Breast cancer is diesis which causes second highest number of death in women. This fact justifies that early detection of breast cancer become the necessary to improve the expectation of life of patients. From last fifteen years, thermography is showing itself as a promising technique for early detection of breast cancer. In this paper, research work carried on the breast thermograms of the existing DMR (Database for Mastology Research) Database on 89 subjects. In 89 subjects 48 subjects were benign whereas 41 subjects were cancerous. The Region of Interest is extracted by proposed algorithm horizontal edge detection followed by Otsu’s thresholding method and morphological operations. Three sets of the features basic: statistical feature, histogram, and GLCM texture features are extracted and the classified by four classifiers MLP, k-NN, SVM, and Ensemble classifier. Highest accuracy of combined features is achieved by 93.53% by ensemble classifier.