8 February 2010 A method for blind estimation of spatially correlated noise characteristics
Author Affiliations +
Proceedings Volume 7532, Image Processing: Algorithms and Systems VIII; 753208 (2010); doi: 10.1117/12.847986
Event: IS&T/SPIE Electronic Imaging, 2010, San Jose, California, United States
Abstract
In design of many image processing methods and algorithms, it is assumed that noise is i.i.d. However, noise in real life images is often spatially correlated and ignoring this fact can lead to certain problems such as reduction of filter efficiency, misdetection of edges, etc. Thus, noise characteristics, namely, variance and spatial spectrum are to be estimated. This should be often done in a blind manner, i.e., for an image at hand and in non-interactive manner. This task is especially complicated if an image is textural. Thus, the goal of this paper is to design a practical approach to blind estimation of noise characteristics and to analyze its performance. The proposed method is based on analysis of data in blocks of fixed size in discrete cosine transform (DCT) domain. This allows further use of the obtained DCT spectrum for denoising and other purposes. This can be especially helpful for multichannel remote sensing (RS) data where interactive processing is problematic and sometimes even impossible.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Nikolay N. Ponomarenko, Vladimir V. Lukin, Karen O. Egiazarian, Jaakko T. Astola, "A method for blind estimation of spatially correlated noise characteristics", Proc. SPIE 7532, Image Processing: Algorithms and Systems VIII, 753208 (8 February 2010); doi: 10.1117/12.847986; https://doi.org/10.1117/12.847986
PROCEEDINGS
12 PAGES


SHARE
KEYWORDS
Image filtering

Image quality

Image analysis

Image processing

Visualization

Remote sensing

Denoising

Back to Top