%0 Journal Article %A Lukin, V. V. %A Abramov, S. K. %A Popov, A. V. %A Eltsov, P. Ye. %A Vozel, Benoit %A Chehdi, Kacem %D 2010 %I Begell House %K blind estimation, noise variance, testing, large database, color images %N 19 %P 1681-1702 %R 10.1615/TelecomRadEng.v69.i19.10 %T A METHOD FOR AUTOMATIC BLIND ESTIMATION OF ADDITIVE NOISE VARIANCE IN DIGITAL IMAGES %U https://www.dl.begellhouse.com/journals/0632a9d54950b268,7ac25b9376db0000,7069233f14801035.html %V 69 %X An automatic method for blind evaluation of additive noise in digital image based on image pre-segmentation, Gaussianity test, and minimal inter-quantile processing of a set of local variance estimates in blocks is proposed. The purposes all aforementioned operations are applied for are discussed. Their joint use allows removing abnormal local estimates that can arise due to image content heterogeneity in blocks and clipping effects that may occur due to several reasons. The proposed method is tested for components of color images in TID2008 database and it is shown to perform accurately enough for most of them. %8 2011-02-10