一种基于子波变换的改进型零树量化编码算法  

An Improved Zerotree Quantization Coding Using Wavelet Transform

作　　者：黄佳庆[1] 翁默颖[1] 

机构地区：[1]华东师范大学电子科学技术系,上海200062

出　　处：《华东师范大学学报（自然科学版）》1999年第3期47-51,共5页Journal of East China Normal University(Natural Science)

摘　　要：该文在图像子波变换的基础上，提出一种改进型零树量化编码算法。该算法保留了原算法的优点，并引入频率优先原则，改进了原算法中存在的不足：当压缩比逐渐增大时，尤其当所处理的图像含高频信息丰富的时候，所恢复图像的低频成份损失较严重，使得重建质量变得很差。实验表明，在相同压缩比条件下，改进算法ＭＳＥ 得到明显减少。尤其对含高频分量丰富的图像，处理效果更为理想。This paper proposes an improved zerotree quantization coding algorithm which is wavelet-based. It keeps good performance of original algorithm. Meanwhile, a principle of ‘Frequency Priority’ is put forward to cut the drawback of original algorithm: when compression ratios are becoming high, the information in low frequency is reduced violently and so the quality of reconstructed image becomes considerably bad, especially when the image includes much high frequency. It can be showed from the experiment that MSE of new algorithm is olwer than the old one. Furthermore, the effect will be ideal when the image has much high-frenquency information.

关 键 词：子波变换 零树量化编码 图像压缩 DCT变换 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] O174.22[自动化与计算机技术—计算机科学与技术]