基于小波子带统计特性的图像量化及质量可分级编码研究

随着数字电视、网上会议和在线点播的出现，越来越多的图像信息需要在网上进行传输，而此时人们不仅要求编码技术具有较好的压缩效果，还要求信息的传输能够满足渐进传输、多质量服务以及图像数据库浏览等一系列的要求。本文提出了一种基于位平面的图像质量可分级编码方法，结合小波子带图像的系数分布特性，利用样本标准差作为量化的依据，对各分辨率级进行分级量化，并且采用了位平面编码和算术编码，使得在解码时可以根据给定解码的数率不同对位平面进行重构以实现图像的质量可分级性。实验结果表明，本文提出的方法实现简单，编解码速度快，在保证具有很好的编码效果的前提下，可以实现图像的质量可分级编码。

The Research of the Quantization Scheme Based on Wavelet Subband Statistical Characteristic and PSNR Scalable Image Coding

With the emergence of digital TV,online meeting and Video-On-Demand,more and more image information need to transmit on the net.People at this moment request not only better compression,but also progressive transmis- sion of image and image database browsing,etc.This paper has put forward a PSNR scalable image coding method based on wavelet statistical characteristic and bit-plane technique.Combining lifting scheme wavelet transform and a- rithmetic coding,the method constructs different bit-planes according to the target bit rates so as to achieve PSNR scal- ability.The experimental results prove,that the proposed method can he implemented simply and achieve high code/ decode spee and good compression quality.Moreover,on the premise of high coding quality,the method can realize im- age's PSNR scalable coding.