Frame-rate adaptation and bi-directional coding for video sequences

In this paper, we present a new method to frame-rate adaptation, which has the goal of improving video quality. Indeed the most popular video coding technique uses a temporal sub-sampled method to decrease the bit rate. Therefore, at the decoding step, the video sequences have a weak frame-rate and a jerky motion. To overcome this problem, it is important to apply a temporal up-sampled technique at the decoding stage to build lost pictures. Currently this problem has no solution, because in spite of all existing methods, none of them offer a robust and really efficient solution. In this paper, we present an original frame rate adaptation method which uses mesh-based motion estimation. This technique provides a solution to perform sub-sampling at the decoding by occlusion area management. Moreover, it leads providing a local intra coding method to accurately recover all occlusion area and improve the robustness and effectiveness. Finally, we propose an easy solution to improve current video coder quality, by adding a B frame processing at the coding and decoding stages.