Model-based Video Coding with Generic 2d and 3d Models

3D extraction from video gives a representation adapted to low bitrate coding and provides enhanced functionalities such as 3D cues for augmented reality and interactive navigation in photorealistic environments. But for degenerated motions of camera, like pure rotation, 3D information can not be retrieved. In this article we propose an original representation based on a hybrid 2D/3D models stream. The idea of this approach is to provide a modelling for all sequences including those with rotations. The sequence is divided into portions and for each one the motion of the camera is identified. Depending on the type of motion a 3D model, a cylindrical mosaic or a spherical mosaic is extracted. They are constructed in order to be suitable for an homogeneous visualization process. Results are shown for synthetic and real video sequences.