摘要
图像的小波变换能同时提供空—频域的局部信息,而矢量量化方法则能取得较高的图像压缩比。在小波变换域实现矢量量化,可望在取得高压缩比的同时,利用小波变换所提供的信息,尽量保证能有较好的恢复图像质量。本文通过在小波变换域适当选择一组交换系数作为矢量,考虑小波变换的频率局部化特性,使用加权均方误差准则,实现一种尺度自适应矢量量化方法。实验结果表明,这种量化方法能取得较明显的效果。
The wavelet transform of images can supply information about both spatial and frequency localizations while vector quantization (VQ) can obtain higher compression ratio. So it is expected that a highen compression ratio and better quality of restored images can be achieved by using the information provided by the wavelet transform and realizing VQ in the wavelet transform domain. In this paper, we realize a scale-adaptive vector quantization (SCAVQ) by using the weighted mean square error (WMSE) selecting a group of tho suitable wavelet transform coefficients as vector as well as considering the property of the freqency localization of the wavelet transform. The experimental result shows that SCAVQ can achieve obvious effect.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1995年第3期188-194,共7页
Pattern Recognition and Artificial Intelligence
关键词
小波变换
矢量量化
图像重建
Wavelet Transform, Vector Quantization. Weighted Mean Square Error, Scale-adaptive Vector Quantization.
