DocumentCode
542587
Title
Potential MSE of color image local filtering in component-wise and vector cases
Author
Uss, M.L. ; Vozel, B. ; Lukin, V.V. ; Chehdi, K.
Author_Institution
TSI2M Lab., Univ. of Rennes 1, Lannion, France
fYear
2011
fDate
23-25 Feb. 2011
Firstpage
91
Lastpage
101
Abstract
In this paper, we investigate the derivation of Cramer-Rao lower bounds on component-wise and vector filtering performance. For two state-of-the-art local (DCT) and non-local (BM3D) filters in both component-wise and vector realizations, filtering output MSEs are compared with their respectively derived CRLBs. The comparative analysis shows good correspondence for images with different content and different noise levels.
Keywords
discrete cosine transforms; filtering theory; image colour analysis; mean square error methods; BM3D filter; Cramer-Rao lower bounds; DCT filter; MSE; color image local filtering; component-wise filtering; vector filtering; Color; Correlation; Databases; Discrete cosine transforms; Filtering theory; Noise; CRLB; color image; component-wise filtering; non-local filtering; potential MSE; vector filtering;
fLanguage
English
Publisher
ieee
Conference_Titel
CAD Systems in Microelectronics (CADSM), 2011 11th International Conference The Experience of Designing and Application of
Conference_Location
Polyana-Svalyava
Print_ISBN
978-1-4577-0042-2
Electronic_ISBN
978-966-2191-17-2
Type
conf
Filename
5744551
Link To Document