A Modified Blind Deconvolution Algorithm for Deblurring of Colored Images

In digital photography, the blur which is caused by motion or camera shake is most common artifact. Due to lack of light or less shutter speed, the image becomes blurry and causes significant degradation in image. The information in the image is affected by this effect. In digital imaging recovering, the unblurred color image from these images is fundamental research problem. There are various types of blur like motion blur, Gaussian blur, etc. Various techniques have been introduced to deblur the images, but specific method is used to deblur specific type of blurring. In this paper, neural networks are used to deblur the images to obtain a sharp one and modified approach in algorithm is introduced in which the image is converted and Sobel color (red, green and blue) filters are applied on the image for better results. The blur kernel is deconvolved out of the blurred image to obtain a sharp image. This approach is better than the baseline approach of using Richardson–Lucy algorithm.