Khalifa, Othman Omran and Dlay, Satnam Singh
(1998)
Wavelets image data compression.
In: Proceedings of IEEE International Symposium on Industrial Electronics, 1998 (ISIE '98), Pretoria, South Africa.
Abstract
The fairt development of multimedia computing
has led to the ibemand of using digital images. The
manipulation, storage and transmission of these images in
their raw form is Cery expensive, it significantly slows the
transmission and bakes storage costly. Many techniques
are now available^ and much effort is being expended in
determining the oF(timum compression technique. Recently,
compression techr/iques using Wavelet Transform (WT)
have received gre:bt attention, because of their promising
compression ratio, ‘flexibility in representing images and its
ability to take into account Human Visual System. In this
paper we combiire a wavelet transform with vector
quantization, usin!! a modified version of LBG algorithm
using Partial Searcih Partial Distortion (PSPD) scheme, for
coding the wavelei1 coefficients to speed up the scheme in
codebook generatic~n and the search required for nearest
neighbour codevdctor of input image. The Wavelet
transform is used to obtain a set of different frequency
subbands of imagte; the image is decomposed using a
pyramidal algorithm architecture. According to Shannon’s
rate distortion theory, the wavelet coefficients are vector
quantized using M\ultiresolution codebooks. The proposed
scheme can save 701 - 80 % of the Vector Quantization (VQ)
encoding time corripared to fully search VQ and reduced
arithmetic compleqity with out sacrificing performance.
Actions (login required)
|
View Item |