An adaptive wavelet-based method for mammographic image enhancement

This paper presents a method for mammographic image denoising and contrast enhancement based on an overcomplete dyadic wavelet transform. The method is aimed at improving the performance of a previous method as well as minimizing user input. The method is based on local non-linear modification of multiscale gradient magnitudes provided by the wavelet transform. Denoising is accomplished by adaptive soft-thresholding and contrast enhancement by a local non-linear gain operator (clipped local range transformation). The processed mammogram is derived by reconstruction from the modified wavelet coefficients. To demonstrate the operation of the algorithm a simulated mammographic phantom with added noise is used. Results of processing a real mammogram are also presented, indicating the potential value of the presented method.