Low Vision X-Ray Images Contrast Enhancement Using Wavelet Transform and Non-Linear Mapping

Abstract

This paper proposes a new method for enhancing the contrast of medical images based on Wavelet Transform. Wavelet transforms offer an efficient representation of the signal, finely tuned to its intrinsic properties involves simple processing techniques in the transform domain, and multi-scale analysis can accomplish remarkable performance and efficiency for many image-processing problems.Wavelet transform is applied so that the intensity values of pixels in gray-level images are decomposed into the approximate component and detail components. The obtained coefficients of the approximate component are normalized before converted by a proposed non-linear grey-level contrast enhancement technique. The non-linearity, with its two degrees of freedom, is more versatile and can produce a more balanced contrast enhancement for low vision medical images. Then, denormalizing results before inverse Wavelet transform application on the converted coefficients, so that enhanced intensity values are obtained. Finally, we found that applying histogram equalization to the result of inverse Wavelet transform promotes the action done in transform domain. The effectiveness of the proposed method is demonstrated experimentally by measuring the contrast ratio and mean ratio for resulted images. The contrast enhancement ratio exceeds 100% of some low vision medical images.