BREAST LESIONS CLASSFICATION USING THE AMALAGATION OF MORPHOLOGICAL AND TEXTURE FEATURES
Sahil Bhusri, Shruti Jain, Jitendra Virmani · 2016
The aim of this paper is to classify the breast lesions using the combination of two feature extraction techniques i.e. morphological features and the texture features. The breast lesions are characterized into two categories Benign and Malignant. Morphological Features computes Area, Perimeter, Convex area, Diameter , Major axis , Minor axis, Extent , Eccentricity, Euler no ,Solidity and Orientation where texture feature /are computed using the statistical features using FOS, GLCM, GLRL, Edge, GLDS, SFM,NGTDM, based statistical feature extraction methods. SVM classifier is extensively used for classification. Using the combination of morphological features and statistical features, the overall classification accuracy of 83.1 % is achieved and the combination of morphological and first order statistics yields the classification accuracy of 89.6%.