Linear Motion Deblurring from Single Images Using Genetic Algorithms

Abstract

One of the key problems of restoring a degraded image from motion blur is the estimation of the unknown linear blur filter from a single blurred input image. Several algorithms have been proposed utilizing image intensity or gradient information. In this paper, we propose an algorithm for restoring the motion-blurred image using Genetic Algorithms. Genetic Algorithms are applied in science and engineering as adaptive algorithms for optimizing practical problems. Certain classes of problem are particularly suited and being tackled effectively with Genetic Algorithm based approach. The direction and the length of the motion blur Point Spread Function (PSF) are used as the parameters of the algorithm. The method assumes a uniform linear camera blur over the image. Experiments on a wide data set of standard images degraded with different directions and blur lengths demonstrate the efficiency of the proposed approach in small blur lengths compared to other algorithms, with a better average Root Mean Squared Error of two values. Experiments also show how ringing artifacts affect the behavior of the algorithm in large blur lengths.