An improved NAS-RIF blind image restoration based on higher order statistics

Author

Song, Chunpeng ; Wang, Hongzhi ; He, Liliang ; Zhou, Mingyue

Author_Institution

Dept. of Comput. Sci. & Eng., Changchun Univ. of Technol.

Volume

2

fYear

2006

fDate

16-20 Nov. 2006

Abstract

The original NAS-RIF blind image restoration algorithm is proposed by D.Kunder and D.Hatzinakos in 1996. This method can restore image when free of noise, but in the presence of noise it can not make greatly well, usually lead to obviously amplification of noise. So based on the defect of the algorithm, a denoise method using higher order statistic (HOS) is proposed. In this method, the blurred image is divided into blocks and third order cumulants is calculated. While image segmentation technology is applied in the every iteration to find the precise object support region in order to improve the performance of the algorithm. Finally, computer simulations show that the proposed method can restore blurred and noised images with whole black, whole white or whole gray background

Keywords

filtering theory; higher order statistics; image denoising; image restoration; image segmentation; recursive filters; blind image restoration algorithm; denoise method; gray background; higher order statistics; image blur; image segmentation technology; noise amplification; nonnegativity and support constraints recursive inverse filter algorithm; Autoregressive processes; Background noise; Degradation; Filters; Higher order statistics; Image processing; Image restoration; Image segmentation; Iterative algorithms; Noise figure;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing, 2006 8th International Conference on

Conference_Location

Beijing

Print_ISBN

0-7803-9736-3

Electronic_ISBN

0-7803-9736-3

Type

conf

DOI

10.1109/ICOSP.2006.345570

Filename

4129051