Contrast Enhancement Using Wavelet Transform and Adaptive Denoising in Mammograms

Breast cancer is a dangerous disease for women worldwide. X-ray mammograms are typically used by radiologists for diagnosis of early-stage breast cancer to reduce the risk of mortality. In many cases they are not easy to be analyzed because mammograms are low contrast and very noisy. This paper proposes improved contrast enhancement and adaptive denoising in mammograms. First, mammograms are decomposed using discrete wavelet transform. Detail subbands are then boosted to increase a global contrast. However it makes noisy mammograms though the contrast improved. The difference between the original mammograms and noisy mammograms is used to identify spatial noise location in the detail subbands. The localized noises are finally suppressed adaptively. The experimental results show higher PSNR than the conventional method while keeping high contrast. Index Terms — Contrast Enhancement, Image Denoising, Wavelet Transform, Mammograms.