Architecture solution for real-time deblurring image/video technique

摘要

This communication addresses the problem of the blur introduced by the out-of-focus defect. We present a new image deblurring technique: it combines a new blur measurement method with the deconvolution algorithm. The proposed technique consists of two steps. The first one allows to estimate the point-spread-function of the optical system by the iterative blind deconvolution algorithm, however the second one consists to recover the image in real-time by the Wiener filter. The main advantage of the proposed technique is that it's no-reference (the blur measurement doesn't need the original image as reference and deblurring image algorithm is blind). The blur measurement method also generalizes and outperforms the blur perception approach that suggested in [10]. Computer simulations are provided in order to illustrate the effectiveness of the proposed approach. A comparison with the method presented in [10] is also performed, emphasizing the good behavior of the proposed blur measurement method.