Detection of Breast Cancer Based on Texture Analysis from Digital Mammograms

In this study, we propose a novel breast cancer detection algorithm based on texture properties of mass area. Proposed method extracts the midpoint of mass area by using AHE (Adaptive Histogram Equalization), and selects the ROT (Region of Interest) in the original image. L1-norm based smoothing filter is then employed to stabilize the mass area, and the form of the mass is determined. Additionally, we measured homogeneity and Ranldet using SVM (Support Vector Machine) to analyze texture properties of the selected mass area. As a result, we observed that the proposed method shows the more stable and outstanding performance for Korean women compared with the existing methods.