Image Deblurring Using Fast Best Kernel Retrieval
Hui‐Yu Huang, Wei-Chang Tsai · 2013
In this paper, we propose a kernel estimation algorithm using fast best kernel retrieval (FBKR) from spectral irregularities to achieve the deblurring. In daily life, when people take a photograph by any kinds of camera equipment, image motion blur caused by camera motion often happens. Camera motion during exposure will produce a blurred image, this kind of blur is often non-liner mode. Motion blur always causes the decline of image quality, as long as the users using the hand-held photography equipment often have a similar experience. Based on this reason, reconstructing a clear image from blurred image is the main objective in this paper. Reconstruction is an ill-pose problem. In current motion blur estimation algorithms, these algorithms usually use the recursive method to estimate motion blur kernel. However, recursive process is quite time-consuming. In order to enhance time-consuming, in this paper, based on iterative phase retrieval algorithm and normalized sparsity measure, we propose a fast and effective blur kernel retrieval algorithm, which can find the best kernel in a short time. Experiments verify that our method can effectively reduce the execution time and obtain the best motion blur kernel to deblur a deblurred image and keep the quality of image deblurring.