文档介绍：pression Using Wavelet Transform and Self-development work
Jung-Hua Wang and Mer-Jiang Gou
Department of Electrical Engineering
National Taiwan Ocean University
Keelung 202, Taiwan, ROC
Email: bQM&&
efficient pressor using wavelet transform and
ABSTRACT
the work. Before training the work, we
In this paper, *e propose a novel method of encoding an first apply the dlscrete Karhunen-Loeve transform to
image without blocky effects. The method mcorporates wavelet coefficients to obtain an optimal set of basis data.
the wavelet transform and a self-development neural Doing so can greatly reduce the quantizatoin errors and
network-Vitality Conservabon (VC) network [ 11 to training time. The block diagram of pression
acheve sigtllficant mprovement in pression system is shown in Fig. 1. anization of hspaper
performance The implementationconsists of three steps. is as follow. Section 2 briefly describes the wavelet
First, the image is posed at different scales using transforms used in th~spaper and a quick review of
wavelet transform to obtain an orthogonal wavelet wavelets in general. In Section 3 we briefly describe the
representation of the image Each band can be discrete Karhunen-Loeve (K-L) "form which is used
subsequently processed in parallel. At the second step, to transform the wavelet coefficients. Section 4
the discre