文档介绍：2006 International Joint Conference on works
Sheraton Vancouver Wall Centre Hotel, Vancouver, BC, Canada
July 16-21, 2006
Improvement of Image Classification with Wavelet and Independent
Component Analysis (ICA) based on a Structured work
Weibao Zou, Yan Li, King Chuen Lo and Zheru Chi
Abstract — Image classification is a challenging problem in Analysis (ICA), feature-based representation of images
organizing a large image database. However, an effective potentially offers an attractive solution to this problem. A
method for such an objective is still under investigation. This number of features can be extracted from the raw image based
paper presents a method based on wavelet and Independent on wavelet [1]. The features of images, such as edges of an
ponent (ICA) for image classification with
object, can be reflected by the wavelet coefficients in low and
adaptive processing of data structures. With wavelet, an image is
posed into low frequency bands and high frequency high bands. Both features or objects and the spatial
bands. An image can be characterized by wavelet coefficients in relationship among them play more important roles in
the form of tree representation. While the histograms of low characterizing image contents because they convey more
frequency wavelet bands are effective in characterizing images, semantic meaning.
the hist