基于ROI图像传输的联合信源信道编码的研究  

Research on joint source channel coding of ROI image transmission

作　　者：席志红[1] 肖春丽[1] 刘晓慧 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《应用科技》2008年第11期11-14,共4页Applied Science and Technology

基　　金：哈尔滨市留学归国人员基金资助项目(2004AFLXJ009)

摘　　要：根据JPEG2000中的一般位移法和最大位移法在处理感兴趣区域时的优点,提出了基于SPIHT算法的感兴趣区域图像压缩编码的新方法,并且充分考虑了SPIHT算法和Turbo码的特性,根据信源编码后数据在重建时的重要程度,利用Turbo码进行了等差错和非等差错保护信道编码.仿真结果表明,在传输全视觉图像时,采用非等差错保护信道编码可以使感兴趣区域的峰值信噪比得到明显提高,并且使感兴趣区域的图像质量远高于背景区域的图像质量.In JPEG2000 standards, both the Generic-scaling-based method and the Maxshift method possess their own advantages in treating regions of interest. Hence, we proposed a new method based on SPIHT for compressed coding of regions of interest. After taking full account of the characteristics of the SPIHT algorithm and the Turbo codes, we used Turbo codes to make the channel coding with unequal and equal error protections according to the importance of coded data in reconstruction. A series of simulations were made. The results showed that the coding with unequal error protection can obviously improve PSNR of ROI, and image quality of ROI is better than that of background region in the process of transmitting full vision images.

关 键 词：感兴趣区域 SPIHT TURBO码 联合信源信道编码 全视觉图像 

分 类 号：TN911.214[电子电信—通信与信息系统]