基于ROI的2D控制图像编码

The Image Coding with Two Dimensions Control Over the ROI

图像是多媒体应用中人们相互之间进行信息交换的重要工具 ,它在 Internet上实时可靠的传输对于多媒体应用的普及具有重要意义。然而 ,当图像在丢包网络如 Internet上传输时 ,接收端有时需要经过长时间的等待才能获得整个码流来重建图像。为此提出了一种基于对象的精细可伸缩性编码方法 ,该方法将对象编码与位平面编码相结合。一方面 ,这种方法具有精细增强图像质量的能力 ;另一方面 ,它又可以通过指定对象优先级和提升位平面来更为灵活地控制用户感兴趣区 (ROI,region of interest)。试验结果证明 。

The image is a vital tool used for communication in the multimedia application. The real time and secure transmission of the image is important to the prevalence of the multimedia application over the Internet. However, when the image transmits over the packets loss channel such as the Internet, the receiver may wait for a long time to get the total Bit stream and then to reconstruct it to the image. This paper presents a new Fine Granular Scalable coding based on the image object, which incorporate the image object coding with the Bit plane coding. On the one hand, this algorithm has the ability granularly enhancing the quality of the image. On the other hand, it has more flexibility in controlling the ROI(region of interest)than the other coding algorithm by setting the priority of the image object and upshifting the bit plane. Firstly, this paper describes the coding framework of this algorithm and then separately discuss the key techniques such as the bit plane coding,selective enhancement,bit stream structure and the controlling strategy. At last, the experiments are given to validate the two features of this algorithm: the fine granular and the flexible control.