Abstract
Image denoising is an important computer vision task, it is also the basis of many other high-level vision tasks. Recently, various image denoising algorithms have proposed, such as BM3D, sparse representation and deep learning, that have achieved impressive performance in AWGN denoising or blind denoising tasks. However, when dealing with Dongba painting (A traditional painting from China’s ethnic minorities), there are problems that Dongba painting have blurred edges and faded colors, it is important to retain its original style when denoising. In this paper, we propose a novel image post-processing technique termed edge-guided directional walk mean smoothing (EDWMS), it can effectively calculate the mean of pixels along different directions then obtain denoised image with sharp edge. Experiments show that the proposed EDWMS outperforms compared algorithms in both edge sharpness and color saturation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have