Multi-view Video Coding via Virtual View Generation

In this paper, a multi-view video coding method via generation of virtual picture sequences is proposed. Pictures are synthesized for the sake of better exploitation of the redundancies between neighbouring views in a multi-view sequence. Pictures are synthesized through a 3D warping method to estimate certain views in a multi-view set. Depth map and associated colour video sequences are used for view generation and tests. H.264/AVC coding standard based MVC draft software is used for coding colour videos and depth maps as well as certain views which are predicted from the virtually generated views. Results for coding these views with the proposed method are compared against the reference H.264/AVC simulcast method under some low delay coding scenarios. The rate-distortion performance of the proposed method outperforms that of the reference method at all bit-rates.