Title :
Factorized Edgeworth-Expanded Gaussians in Topographicmap Formation
Author :
Van Hulle, Marc M.
Author_Institution :
Lab. voor Neuro- en Psychofysiologie, Katholieke Univ. Leuven, Leuven
Abstract :
We factorize the Edgeworth-expanded Gaussian kernel so that the number of parameters to estimate increases linearly with the input dimension, rather than polynomially. We develop this concept for kernel-based topographic map formation and use it for image feature extraction.
Keywords :
Gaussian processes; feature extraction; image processing; parameter estimation; factorized Edgeworth-expanded Gaussian kernel; image feature extraction; parameter estimation; topographic map formation; Clustering algorithms; Feature extraction; Gaussian processes; Independent component analysis; Kernel; Laboratories; Neurons; Parameter estimation; Polynomials; Psychology;
Conference_Titel :
Machine Learning for Signal Processing, 2006. Proceedings of the 2006 16th IEEE Signal Processing Society Workshop on
Conference_Location :
Arlington, VA
Print_ISBN :
1-4244-0656-0
Electronic_ISBN :
1551-2541
DOI :
10.1109/MLSP.2006.275549
