Optimization of Lucy-Richardson Algorithm Using Modified Tikhonov Regularization for Image Deblurring

Abstract

In this paper, we propose the method to optimize the iteration number of the Lucy-Richardson algorithm for image deblurring. This technique based on the modified Tikhonov regularization which composed of 2 parts which are designed for measuring the image similarity and noise enhancement due to the deblurring process. The regularization parameter will be used to control the desired deblurred image. Several sizes of the Gaussian blur kernel are applied for generating the degraded image in the simulation experiment. The Peak Signal to Noise Ratio (PSNR) metric is used to measure the deblurring performance. The results show that this method can be used to estimate the optimal iteration number and it also gives the PSNR value higher than the default Lucy-Richardson method and regularized filter all sizes of the experimental blur kernel. Moreover, it also tolerance to the deviated blur kernel especially it smaller than exact blur kernel.