Wavelets image data compression

• Main Authors: Khalifa, Othman Omran, Dlay, Satnam Singh
• Format: Conference or Workshop Item
• Language: English
• Published: 1998
• Subjects: T Technology (General)
• Online Access: http://irep.iium.edu.my/5811/1/00711556.pdf; http://irep.iium.edu.my/5811/; http://dx.doi.org/10.1109/ISIE.1998.711556
• Institution: Universiti Islam Antarabangsa Malaysia

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.