Monte-Carlo SURE for Choosing Regularization Parameters in Image Deblurring

Yubing Han (Nanjing University of Science and Technology)
Kelan Wang (Nanjing University of Science and Technology)
Mengna Xu (Nanjing University of Science and Technology)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 11, No 6: June 2013

Parameter choice is crucial to regularization-based image deblurring. In this paper, a Monte Carlo method is used to approximate the optimal regularization parameter in the sense of Stein’s unbiased risk estimate (SURE) which has been applied to image deblurring. The proposed algorithm is suitable for the exact deblurring functions as well as those of not being expressed analytically. We justify our claims by presenting experimental results for SURE-based optimization with two different regularization algorithms of Tikhonov and total variation regularization. Experiment results show the validity of the proposed algorithm, which has similar performance with the minimum MSE. DOI: http://dx.doi.org/10.11591/telkomnika.v11i6.2675