一种基于RCPT码的渐进图像传输方法

Progressive Image Transmission Using RCPT Codes

将等级树中的集分割编码器输出的码流分组,根据输出的分组数据对重建图像重要性的不同,传输前采用纠错能力不同的RCPT码保护。在总的码率限制下,采用遗传算法从有限码率集中选择每组数据的RCPT码码率,使接收端恢复的图像失真最小。将等误差保护码率作为遗传算法的初始种群,减少了最优码率集的搜索时间。在二进制对称信道上的仿真结果表明,该渐进图像传输方法实现简单,在高误码率信道条件下恢复图像仍能得到很高的峰值信噪比。

The bit stream from the set partitioning in hierarchical trees coder is divided into packets, which are protected by RCPT codes of different error-correcting capability before transmission according to their effect to the reconstruction of image. The genetic algorithms are developed for selecting optimal RCPT codes rates from the finite rates set in order to minimize the end-to-end distortion. The time for searching the optimal rate set is reduced when using the channel code rates for equal error protection as the initial condition of the genetic algorithms. Simulation results in binary symmetric channel show that this transmission scheme for progressive image is easy to perform and able to achieve high peak signal-to-noise ratio even in high bit error rate channel.