Total Variation and Alternate Direction Method for Deblurring of Digital Images

Images captured using smartphones and video cameras are recorded and can be used anywhere and at any time. While taking a quick shot or while capturing the moving objects, it may lead to the motion blurred images. In order to recover the sharp images from motion blurred images, blind motion deblurring can be used. Motion deblurring can be done by knowing both edge and non-edge of motion blurred images. Edge and non-edge are the two methods used in total variation and alternate direction method for deblurring of digital images. Step edges can be predicted and detected by using edge-specific method. In non-edge method, it explores various image statistics, such as the prior distributions and it is sensitive to statistical variation over different images. Both methods are used in large dataset images, but it fails extremely in simple images. To overcome this problem, total variation (TV) based regularization method is used which is followed by an iteratively reweighted algorithm based on alternating direction method. To get higher results, LSED prediction—based technique is employed, which first of all restores sharp edges and then uses them to estimate initial kernel that traps the optimization of local minimum corresponding to sharp images.