Title :
Linear Transformation Technology for Image Feature Drop Dimension
Author :
Xiao, Feng ; Zhou, Mingyuan ; Geng, Guohua
Author_Institution :
Sch. of Inf. Sci. & Technol., Northwest Univ., Xi´´an, China
Abstract :
Linear Transformation Technology can eliminate component relevance of image high dimension feature vectors and drop dimension of feature vectors, then extract image features effectively. This paper analyses and discusses the methods of PCA (Prinapal Component Analysis), ICA (Independent Component Analysis), and SVD (Singular Value Decomposition) based On Linear Transformation Technology. The methods of PCA and SVD can eliminate 2-order relevance between feature vectors and ICA can eliminate high-order relevance between inputed feature vectors.
Keywords :
feature extraction; independent component analysis; principal component analysis; singular value decomposition; ICA; PCA; SVD; image feature drop dimension; image feature extraction; image high dimension feature vector; independent component analysis; linear transformation technology; principal component analysis; singular value decomposition; Data mining; Educational institutions; Feature extraction; Image coding; Principal component analysis; Semantics; Vectors; independent component analysis; linear transformation; prinapal component analysis; singular value decomposition;
Conference_Titel :
Knowledge Acquisition and Modeling (KAM), 2011 Fourth International Symposium on
Conference_Location :
Sanya
Print_ISBN :
978-1-4577-1788-8
DOI :
10.1109/KAM.2011.95
