基于低频预测的图像多描述编码及其差错隐藏  

Low frequency prediction-based multiple description image coding and error concealment

作　　者：王红艳[1] 金炜东[2] 

机构地区：[1]西南交通大学信息科学与技术学院,成都610031 [2]西南交通大学电气工程学院,成都610031

出　　处：《计算机工程与应用》2013年第18期159-163,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60971103)

摘　　要：为了提高图像、视频在丢包或误码环境下的传输性能,提出了基于低频预测的图像多描述编码方法。根据小波系数的独特特点将原始图像小波变换后的系数分为低频,次高频,高频三部分,对各部分进行不同方式的分组;采用低频预测的预处理和后处理方案,实现描述间的冗余插入;实现了差错隐藏算法以提高传输的鲁棒性。仿真结果表明,系统的抗误码能力大为提高,且可提供灵活冗余插入以适应各种网络传输状况。To improve the transmission performance of images or video in the environment of packet loss or bit error,multiple description image coding method based on low-frequency prediction is proposed.By the unique characteristics of wavelet coefficient,wavelet transform coefficients of the original image are divided into three-part of low-frequency,minor-high frequency,high frequency,and these sections are grouped in different ways.Pre-processing and post-processing program of low frequency prediction are used to achieve redundant insertion.Error concealment algorithm to improve the robustness of transmission is realized.Simulation results have shown that this new algorithm greatly improves systems-resilience capabilities,provides flexible redundant insertion to adapt to network conditions.

关 键 词：图像编码 多描述编码 小波系数 预处理 差错隐藏 

分 类 号：TN919.81[电子电信—通信与信息系统]