Breast Cancer Diagnosis in Digital Mammogram Using Steerable Pyramid

Abstract

This article is to explore the usefulness of steerable pyramid as feature extraction for breast cancer diagnosis in digital mammogram. The proposed approach is compared to some related feature extraction methods as Discrete Wavelet Transform (DWT), linear discriminant analysis (LDA) and principal component analysis (PCA). K-nearest neighbor, Support Vector Machine and Naive Bayesian is used separately to construct a supervised classifier. Experimental results on Digital Database Screening Mammography breast cancer database provides a better representation of the class information, and obtains much higher recognition accuracies compared with well established methods in literature.