A hybrid approach for image denoising using explicit guided filter and collaborative wiener filter

In this paper, we propose an image denoising strategy which combines the edge-preserving property of an explicit filter, known as the guided filter, with the de-blurring property of a collaborative wiener filter. Removing noise from an image without losing the main image features such as object edges, corners, and other fine structures, while minimizing the blurring effects in the denoised image is the main objective of our proposed strategy. The guided filter is used in the preprocessing step and it takes into account the content from a reference image and transfers the structure of this image to the output, i.e. preprocessed image. The noisy input image is then processed by the wiener filter, which also involves the use of this preprocessed output image for collaborative filtering. The experimental results of our proposed approach show that not only the edges and other refined features in the input noisy image are preserved to a great extent, but also the extent of blurring is reduced to a satisfactory level. Our results, when compared to those of Guided filter, demonstrate that we are able to achieve a better denoising performance than this approach, regardless of the type of image used.