Mammograms contrast enhancement based on background homogenization

Abstract

Breast cancer is one of the most pervasive women diseases. Mammography can greatly facilitate the early detection of breast cancer. The low quality of the mammogram undermines the accuracy of diagnosis. Image enhancement techniques are often used to enhance the mammogram contrast and assist the radiologists to detect the cancer. Histogram Equalization (HE) is one of the most widely used methods for enhancement. However, after HE processing, dark region remains unfading. The edge of mammogram which provides important information for diagnosing usually belongs to dark regions. To increase the luminance range, we proposed a new method for contrast enhancement by compressing luminance and applying HE. The intensity depends on the mean values of both small windows and the original image. The proposed algorithm showed promising performance when applied on several mammogram.