摘要
从人眼视觉特性的角度研究基于小波变换的图像压缩算法．分析了图像经小波变换后得到的子带图像的视觉特性和统计特性，对传统的矢量量化算法给予了改进，提出了一种基于小波变换、符合人眼视觉特性的矢量量化－－－Ｗａｌｓｈ－Ｈａｄａｍａｒｄ变换域下的矢量量化（ＷＨＴ－ＶＱ）算法．设计多分辨率码书，为不同分辨率下、不同方向上的高频子带选择了相应的Ｗａｌｓｈ－Ｈａｄａｍａｒｄ基图像．重建图像的主观视觉效果良好，而且具有较高的压缩倍率．
An algorithm for image compression based on wavelet transform was presented in the view of human visual system. The visual characteristics and statistical characteristics of wavelet subimages were studied. Based on the conventional VQ algorithm (such as LGB algorithm),a modified VQ algorithm——WHT VQ algorithm was put forward.In this paper,a kind of selection method for WHT basis images was given corresponding to detail subimages of different directions and resolutions. Results showed that the decoded images have good quality as well as high compression ratio.
出处
《北京理工大学学报》
EI
CAS
CSCD
1997年第6期734-738,共5页
Transactions of Beijing Institute of Technology
基金
电科院电子科技预研基金
关键词
图像压缩
小波变换
矢量量化
数字图像编码
image compression
wavelet transform
visual characteristics
vector quantization
