摘要
基于数字图像压缩的可能性与必要性,本文用信息论观点讨论了数字图像压缩的方法。并分别采用差分脉冲编码调制、离散余弦变换、离散小波变换等3种方法,对Lena(512×512)图片进行处理,统计处理后的系数的概率分布、标准差、信息熵,以及增大量化阈值后0系数所占的比率,来研究他们用于数字图像压缩的性能。同时,还给出了每种方法的压缩重建的误差分析。并做了详细地比较,分析了导致他们性能差异的原因,这对于进一步学习和寻找更好的数字图像压缩方法具有一定的指导意义。
For the possibility and necessity of digital image compression, this paper has discussed the principle of three usual algorithms applied to digital image compression. Including Differential Pusle Code Modulation（DPCM）, Discrete Cosine Transform（DCT）, Discrete Wavelet Transform（DWT）. What＇s more, for researching their performance on image compression, it has given out the probability distributing of coefficient, standard difference, information entropy, mean squared error, and the percentage of zero coefficient after processing the picture Lena（512×512） with them on Matlab. It has a comparison of these three algorithms in detail,and has analysed the cause of leading to their different performance, this contributes to further learning and searching the better algorithms for digital image compression.
出处
《现代电子技术》
2005年第16期69-72,75,共5页
Modern Electronics Technique
