摘要
不等差错保护(UEP)在数据传输中具有十分重要的作用。在多媒体传输中,如JPEG2000图像的传输,数据通常依据重要性进行分类。重要的数据应得到更好的保护,即在传输中具有更低的误比特率(BER)。本文基于误比特率分布,深入研究了Turbo码中不同删截方案对不等差错保护特性的影响。基于“逐渐稀疏”的思路,提出了一种能够产生强不等差错保护的Turbo码删截方案。和其他删截方案相比,本方案可以在对重要数据提供强不等差错保护的同时,尽量减小不重要数据的误比特率。将此方案应用到JPEG2000图像传输中可以看到,和其他方案相比,峰值信噪比(PSNR)有0.5-0.9dB的改善。
Unequal Error Protection(UEP)is a powerful framework for protecting scalable bitstreams against bit errors.In multimedia transmission systems,such as the transmission of JPEG2000,data usually is divided into levels according to their different importance.In these cases,UEP techniques can guarantee lower bit error rate(BER)for the more important classes.In this paper we deeply study the affection of puncturing pattern to UEP based on BER distribution.Then a new scheme of puncturing patterns for turbo codes with strong UEP is proposed.In the new designing scheme,an idea of“gradually sparsing”is employed.Compared with other puncturing schemes,this scheme can provide stronger error protection for some bits without apparent degradation for other bits.simulations show that the advantage of the proposed puncturing scheme over other puncturing schemes presented in references is obvious,about 0.5-0.9 dB in decoded image Peak Signal-to-Noise Ratio(PSNR).
作者
邵霞
王超
SHAO Xia;WANG Chao(North China University of Water Resources and Electric Power,Zhengzhou Henan,450045,China)
出处
《数码设计》
2018年第2期15-17,共3页
Peak Data Science
