Restoration of discrepancy iterative images using mechanism for evaluating noise

被引：0

作者：

Tang, Xinjian ^{[1
]}

Liu, Jian ^{[1
]}

Tian, Jinwen ^{[1
]}

机构：

[1] Institute of Patters Recognition and Artificial Intelligence, Huazhong University of Science and Technology, Wuhan 430074, China

来源：

Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition) | 2007年 / 35卷 / 11期

关键词：

Algorithms - Convergence of numerical methods - Estimation - Image processing - Iterative methods - Wavelet transforms;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

Morozov discrepancy equation was investigated and formulated by Tikhonov regularization solution that was derived from the degraded image model. The image noise, as the knowledge a prior, was estimated by employing wavelet transform and the equation was modified. A novel method that seeks the optimal regularization parameter by iteration with the improved Morozov discrepancy equation was proposed. The experimental results show that under a lack of the noise information of the observed image the iterative convergence speed and restoration effect of the proposed method were obviously superior to the Van Cittert iterative algorithm. The iterative speed of the proposed algorithm was thrice to quintuple as fast as the Van iteration and PSNR (peak signal-noise rate) was 1 to 5 dB bigger than Van method. As a conclusion the proposed image restoration method is applicable and promising.

引用

页码：59 / 61