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

We present a new method for frame-rate adaptation, with the goal of improving video quality. The most popular video coding technique uses a temporal subsampled 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 upsampled technique at the decoding stage to build lost pictures. Currently this problem has no solution, because none of the existing methods offer a robust and really efficient solution. We present an original frame rate adaptation method which uses mesh-based motion estimation. This technique provides a solution to perform subsampling during decoding by occlusion area management. Moreover, it leads to providing a local intra coding method to recover accurately all the 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.