Color image deblurring and denoising by using bilateral sharpening

Abstract

Images can be affected by noise and blur, which can cause quantitative bias and segmentation problems. In this work, we propose a new method to enhance the signal-to-noise ratio of 2D degraded color images while sharpening blurry areas. The method consists of the original combination of indirect bilateral filter and sharpens term based on local structural analysis. The structure tensor brings the local orientation of multi-channel image features. The proposed method reduces the noise inside regions while restoring contrast between regions. We present results on synthetic images, which led to distinct improvements of merit figures over two methods from the literature.