<title>Classified wavelet transform coding of images using vector quantization</title>

Abstract

The discret wavelet transform (DWT) has recently emerged as a powerful technique for image compression in conjunction with a variety of quantization schemes. In this paper, a new image coding scheme -classified wavelet transform/ vector quantization (DWT/CVQ) -is proposed to efficiently exploit correlation among different DWT layers aiming to improve its performance. In this scheme, DWT coefficients are rearranged to form the small blocks, which are composed of the corresponding coefficients from all the subbands. The block matrices are classified into four classes depending on the directional activities, i.e., energy distribution along each direction. These are further divided adaptively into subvectors depending on the DWT coefficient statistics as this allows efficient distribution of bits. The subvectors are then vector quantized. Simulation results show that under this technique the reconstruction images preserve the detail and structure in a subjective sense compared to other approaches at a bit rate of 0.3 bit/pel.

Wavelet‐based adaptive vector quantization for still‐image coding

“…Bit rates (BPP) PSNR (dB) Chen et al (1995) 0.132 27.70 Antonini et al (1992) 0.150 28.01 Antonini et al (1992) 0.210 29.11 Huh et al (1995) 0.200 29.72 Paek and Kim (2000) 0.255 33.12 Paek and Kim (2000) 0 coefficient vectors for each level individually. For a four-level wavelet-transformed image, the input vectors composed are 2 ϫ 2, 2 ϫ 2, and 4 ϫ 4 blocks for level 4, level 3, and level 2, respectively.…”

“…The vectors are composed by grouping the coefficients in the same orientation from each level, with a different scanning order for each different orientation. WT theory has three key features: (1) Wavelet coefficients in horizontal orientation contain the vertical edge information; (2) the coefficients in vertical orientation contain the horizontal edge information; and (3) the coefficients in diagonal orientation do not have regular distribution (Huh et al, 1995;Mallat, 1989a, b;Shapiro, 1993). To keep the similarity between the vectors in the same orientation, three different scanning orders, as shown in Fig.…”

“…The classified VQ in the wavelet domain reduces the edge distortion and the computation complexity (Nasrabadi and King, 1988). An alternative classified VQ (Huh et al, 1995) considers the wavelet coefficients in three directions at the same time. The coefficients are classified into four classes: class H (horizontal direction), class V (vertical direction), class D (diagonal direction), and class L (low-frequency component).…”

“…Many wavelet-transformed coding schemes have been proposed (Huh et al, 1995;Shapiro, 1993). Because of the inherent multiresolution features in WT, not only intrascale correlation between the coefficients within a scale but also interscale correlation between different scales are employed to design a high-performance coding scheme in WT.…”

Fast solution of problems with multiple load cases by using wavelet‐compressed boundary element matrices

“…Although very e cient, these techniques require completely new numerical formulations and related computer programs. Another alternative exploited in this paper is to apply wavelet transforms, a technique widely used in signal and image compression [9,10].…”