Optimised Motion Estimation for Robust Video Coding in Packet Loss Environment
Fuzheng Yang, Shuai Wan, Ebroul Izquierdo · 2007
An improved motion estimation method for robust video coding is proposed. It is designed to enhance packet loss resilience in lossy environments. The proposed motion estimation is based on an end-to-end rate-distortion optimisation framework. Initially, the end-to-end reconstructed distortion is efficiently estimated considering the distortion caused by quantisation, error propagation, and error concealment. Then, the total bit rate for coding the residual is estimated using a quadratic rate-distortion model. The results are incorporated into the Lagrangian optimisation during motion estimation. Optimised motion vectors are selected in the sense of coding efficiency and robustness. A comparative evaluation with conventional robust video coding techniques was conducted. The experimental results demonstrate a superior performance of the proposed method.