A Variational Approach to Image Despeckling under Varied Blur
Ramesh Kumar Thakur, Suman Kumar Maji · 2020
Restoring blurred and speckle corrupted images have been a traditional problem in the area of image processing. In this paper, we introduce a new variational Bayesian approach for deblurring and despeckling of images corrupted by multilevel noise and different types of blur. We introduce a L1norm based data-fidelity term and coupled with Total Variational (TV) regularization filters in our optimization problem. Such an approach, as is evident through experimental result, is capable of retaining finer details in the image while producing high quality despeckled and deblurred image. Comparison with existing state-of-the-art shows the high-quality visual and quantitative results, justifying the superiority of the proposed approach.