Object-Based Stereoscopic Conversion of MPEG-4 Encoded Data

摘要: Stereoscopic conversion of two-dimensional (2-D) video is considered in object-based approach that independently processes each object. Our works extend the previous frame-based stereoscopic MPEG-1 and 2 to MPEG-4. In MPEG-4, image composed a background object primary object(s). first step, camera motion type determined for generating image. For this, vectors are utilized. The generation makes use current As well, uses its horizontally-shifted version avoid possible vertical parallax might happen. second two objects combined generate verified experiments performed on MPEG-4 sequences, can cope with has been difficult problem deal approach.