The Performance Evaluation of the Breast Microcalcification CAD System Based on DWT, SNE AND SVM

Abstract

Mammogram is measured the most consistent method for early detection of breast cancer. Computer-aided diagnosis system is also able to support radiologist to detect abnormalities earlier and more rapidly. In this paper the performance evaluation of the computer aided diagnostic system for the classification of microcalcification in digital mammogram based on Discrete Wavelet Transform (DWT), Stochastic Neighbor Embedding (SNE) and the Support Vector Machine (SVM) is presented. This proposed system classifies the mammogram images into normal or abnormal, and the abnormal severity into benign or malignant. Mammography Image Analysis society (MIAS) database is used to evaluate the proposed system. The average classification rate achieved is very satisfied.