摘要
从小波滤波器的多相位矩阵出发,利用因子分解方法实现了小波变换由传统卷积模式转变到提升体制,进而在此基础上实现了整数小波变换,并利用零树编码算法对变换后的系数进行了重新组织。通过综合研究它们各自的特点,设计了同时具有高效压缩性和嵌入式码流特性的图像编解码系统,试验表明该系统可以更好地满足图像信息在存储、网络浏览以及传输方面的需求。
In this paper,the polyphase matrix of wavelet filters is factorized into a sequence of triangular matrixes and a diagonal matrix. The wavelet transform is implemented with lifting scheme instead of the traditional convolution and the integer to integer wavelet transform is achieved on the basis of lifting scheme. The wavelet coefficients are reconstructed by the method of embedded zerotree encoding. The image eodec is designed by combining the characteristics of integer to integer wavelet transform and embedded code stream. The experimental results show that the proposed algorithm meets the needs of storage, Internet browsing and transmission of image.
出处
《计算机应用研究》
CSCD
北大核心
2005年第9期159-161,共3页
Application Research of Computers
基金
国家自然科学基金重点项目(60272022)
关键词
整数小波变换
零树算法
图像编解码系统
Integer to Integer Wavelet Transform
Zerotree Encoding
Image Codec
