摘要
随着数字电视和网上会议的出现,越来越多的视频信息需要在网上进行传输,而视频压缩效率的提高在很大程度上取决于如何减少用于传输运动补偿预测误差的信息量,即如何使视频帧差图像满足渐进传输、多质量服务以及图像数据库浏览等一系列的要求。首先,结合视频帧差图像的统计特性,提出了一种帧差图像小波系数的双重量化方法,在此基础上提出了一种基于位平面的帧差图像质量可分级编码方案,在解码时能够根据给定解码的数率对位平面进行重构,并实现帧差图像的质量可分级解码。实验结果证明,提出的方法编码效果优于传统的SPIHT编码算法,特别在低比特率下具有很好的编码效果。
With the appearances of digital TV and online meeting, more and more image information needs to transmit on the net. The improvement of video coding depends on how to reduce the use of motion compensation prediction residuals(MCPR), which also means the MCPR image should have the ability of progressive transmission, multi-quality services and image database browsing, etc.. In this paper, a double-stage quantization method is proposed combining with the statistica property of MPCR image. Then a scalable coding method for frame difference image quality based on bit-plane technique is proposed. The method constructs different bit-planes according to the target bit rates so as to achieve the scalable decoding for frame difference image quality. Simulation results prove that the compression quality of the method is superior to the SPITH algorithm, especially at the very low bit rates.
出处
《微型电脑应用》
2015年第12期5-10,4,共6页
Microcomputer Applications
基金
辽宁省教育厅科学研究一般项目(L2011192)
大连市科学技术基金计划项目(2013J21DW027)
关键词
帧差图像
统计特性
可分级编码
提升方案小波
Prediction Error Image
Statistical Property
Scalable Coding
Frame Difference Image
Lifting Scheme Wavelet
