Two Class Classification of Breast Lesions using Statistical and Transform Domain features

Abstract

In breast cancer the breast lesions are differentiated into two classes: Benign and Malignant. In this paper Computer- Aided Detection (CAD) system is designed for detecting lumps which may indicate presence of breast cancer. This paper presents the  classification of breast ultrasound images using Statistical and Transform domain feature extraction techniques were data is partitioned by hold-out method and classified using Support Vector Machine (SVM) classifier. The aim of this paper is to achieve higher accuracy in classification of the lesions. SVM trains a model that assigns unseen new objects into a specific category. The best obtained result out of all the features used is calculated using Fourier Power Spectrum (FPS) with 91.5 % overall accuracy and 81.2 % and 92.8 % individual class accuracies for benign and malignant class respectively .