PDE-based enhancement of color images in RGB space.

Abstract

A novel method for color image enhancement is proposed as an extension of the scalar-diffusion-shock-filter coupling model, where noisy and blurred images are denoised and sharpened. The proposed model is based on using the single vectors of the gradient magnitude and the second derivatives as a manner to relate different color components of the image. This model can be viewed as a generalization of the Bettahar-Stambouli filter to multivalued images. The proposed algorithm is more efficient than the mentioned filter and some previous works at color images denoising and deblurring without creating false colors.