Enhancement of multi-valued images using PDE coupling

Abstract

In this paper, we present a new model for the enhancement of noisy and blurred multi-valued images. The proposed model is based on using single vectors of the gradient magnitude and the second derivatives as a manner to relate different colour components of the image. This model can be viewed as a generalization of Bettahar-Stambouli filter to multi-valued images. The proposed algorithm is more efficient than the mentioned filter and some previous works at colour images denoising and deblurring without creating false colours.