【作者】 覃剑；

【导师】 曾孝平；

【作者基本信息】 重庆大学 ， 信号与信息处理， 2003， 硕士

【摘要】 图像在信息获取中具有重要意义，但由于数据量大的特点，给传输和储存带来很大不便。传统的图像压缩方法主要从信息论的角度出发，没有考虑视觉特性，在压缩率和恢复图像的主观质量上不能满足人们的要求。本文主要研究如何利用图像小波变换后系数间的相关性实现小波系数的分级编码和传输，从而实现图像的渐近传输技术。 首先，由于小波分析由于具有多分辨或多尺度特性，在图像处理方面得到广泛应用。提升技术是构造小波和实现小波变换的一种新方法，并可以实现从整数到整数的变换，因此可以成为图像无损压缩的新手段。 其次，由于不同用户对同一图像的质量要求不同，因此有必要在编解码时，使图像的恢复能够根据用户要求，做到从有损到无损的过渡，即图像的渐近编码。 本文主要利用提升算法实现图像的整数小波变换，然后对几种经典的嵌入式图像编码算法进行了研究和比较，在对经过小波变换的图像系数进行统计分析后，提出一种对EZW(Embedded Zerotree Wavelets)算法进行改进的嵌入式图像压缩编码算法。该算法旨在提高图像无损压缩的压缩率，并可用于在进行图像有损压缩时提高恢复图像的主观质量。经过仿真表明，该算法在压缩率和恢复图像的主观质量上优于EZW算法。更多还原

【Abstract】 Image is important for us to obtain information. But it is difficult to save and transmit because of abundant data. Tradition image compression method mostly consider information theory instead of vision character. So it is not satisfy the people’ s need on compression ratio and revert quality. This paper mainly study how to realize wavelet coefficient layered coding and transmission and image data progressive transmission using the correlation of the transformed image.First, wavelet was widely used in image processing due to its multiresolution or multiscale analysis property. Lifting scheme is a new method for constructing wavelets and performing wavelet transform. At the same time, it can realize transforms that map integers to integers. So lifting scheme will become a new method in lossless image compression.Secondly, image data should be transmitted from loss to lossless, named progressive transmission , because different people require different image quality.This paper mainly realize integer wavelet transform using lifting scheme, then study and compare several embedded coding arithmetic. Base on analyzing the image coefficient, we propose an improved EZW(Embedded Zerotree Wavelets) arithmetic. Jt aimed at increasing compression ratio in lossless image compression and subjective quality in loss image compression. Experiment shows that this arithmetic is superior to EZW arithmetic in compression ratio and subjective quality.更多还原