A slice-length selection algorithm based on the frame-level RDO

Abstract

With the rapid development of video communications between heterogeneous networks and video surveillance services, scalable video coding (SVC) technology is proposed to solve the fluctuation of network bandwidth and become the current research focus. Because bit error rate in wireless channel is very high, and it is the serious impact for video quality, the error concealment algorithm research of the scalable video has begun to receive extensive attention. The paper studies the spatial error concealment algorithm for SVC; it will establish frame-level rate-distortion optimization (RDO) model based on scalable video coding framework. While coding the each SVC layer, and planning a flexible slice-packaged strategy for each frame image to establish re-synchronization mechanism in order to prevent the spatial error propagation, as the same bit error rate, the decoded video image quality will improve 0.5~1.5dB. Research findings of the paper will satisfy the requirements of wireless video communications and mobile video surveillance, and so on.