Contrast enhancement mammograms using denoising in wavelet coefficients

Abstract

Contrast enhancement in x-ray mammograms is important in improvement of radiologists' reading and interpretation. In many cases, it is difficult to discern signs of breast cancer because mammograms are low contrast and very noisy. This paper proposes improved contrast enhancement method in mammograms using denoising in wavelet coefficients. First, mammograms are decomposed by wavelet transform. Second, the ratio between approximate subband and detail subband as signal-to-noise ratio is computed. Finally, detail subbands which have the ratio lower than the criteria are boosted to improve mammograms contrast while detail subbands having the ratio higher than the criteria are set to zero for noise suppression in the same time. Contrast measure and peak-signalto-noise ratio are used to evaluate the performance of the proposed method. The experimental results show higher contrast 17.05% than the conventional method while a bit different in peak-signal-to-noise ratio.