Modified Newton Integration Algorithm With Noise Tolerance for Image Deblurring
Shan Liao, Haoen Huang, Jiayong Liu, Xiuchun Xiao, Xiaoyang Li, Shubin Li · IEEE Transactions on Computational Imaging · 2021
This paper addresses image deblurring problem, which consists of recovering a rather high resolution image from its blurred and noisy observation version. Different from the existing solving methods, influence of noise perturbation is considered when modeling inverse solution process. To remedy shortcomings that the traditional and current computing methods have encountered under a noisy environment, a modified Newton integration (MNI) algorithm is presented from the control perspective. Besides, through the corresponding theoretical analyses and proofs, the proposed MNI algorithm is confirmed to handle image deblurring problem with noise-tolerance ability and fast convergence performance. Finally, multiple simulative experiments are performed to verify excellent image deblurring ability and noise-tolerance advantage of proposed MNI algorithm compared among Newton-Raphson iterative (NRI) algorithm, gradient descend dynamic (GDD) algorithm, and discrete time derivative (DTD) algorithm.