适于感兴趣区域编码的多级树分裂算法研究

Research of SPIHT Algorithm Adopted for ROI Image Coding

结合JPEG2000中最大平移法和基于尺度法对感兴趣区域处理的优点,对多级树分裂算法(SPIHT)进行了改进,实现了一种适合感兴趣区域的图像编码方法。与JPEG2000中最大平移法和基于尺度法相比,该方法在低比特率的情况下,不但可以传输ROI区域的图像,还可以传输部分背景区域图像,并且在ROI区域传输完毕的情况下,不再传输ROI上的0码,节省了码流,进一步提高了编码效率。

On the basis of SPIHT algorithm,a novel method adapted to ROI Image Coding is proposed, combining with the advantages of max shift method and generic scaling based method for processing of ROI in JPEG2000. Compared with the two methods in JPEG2000, the method introduced in the article can transmit not only ROI image, but also part of background region image. After the ROI image is transferred thoroughly, the zero code in the ROI can be interrupted,which saves the code stream and improves the coding efficiency.