DocumentCode :
2231716
Title :
Modified ICA algorithms for finding optimal transforms in transform coding
Author :
Narozny, Michel ; Barret, Michel ; Pham, Dinh-Tuan ; Akam-Bita, Isidore Paul
Author_Institution :
Equipe Syst. de Traitement du Signal, France
fYear :
2005
fDate :
15-17 Sept. 2005
Firstpage :
111
Lastpage :
116
Abstract :
In this paper we present two new algorithms that compute the linear optimal transform in high-rate transform coding, for non Gaussian data. One algorithm computes the optimal orthogonal transform, and the other the optimal linear transform. Comparison of the performances in high-rate transform coding between the classical Karhunen-Loeve Transform (KLT) and the transforms returned by the new algorithms are given. On synthetic data, the transforms given by the new algorithms perform significantly better that the KLT, however on real data all the transforms, included KLT, give roughly the same coding gain.
Keywords :
Karhunen-Loeve transforms; independent component analysis; transform coding; Karhunen-Loeve Transform; modified ICA algorithms; non Gaussian data; optimal linear transform; optimal orthogonal transform; transform coding; Context modeling; Discrete cosine transforms; Discrete transforms; Discrete wavelet transforms; Image coding; Independent component analysis; Karhunen-Loeve transforms; Quantization; Transform coding; Vectors;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Image and Signal Processing and Analysis, 2005. ISPA 2005. Proceedings of the 4th International Symposium on
ISSN :
1845-5921
Print_ISBN :
953-184-089-X
Type :
conf
DOI :
10.1109/ISPA.2005.195393
Filename :
1521272
Link To Document :
بازگشت