Stereoscopic conversion of two-dimensional movie encoded in MPEG-2

We present an algorithm for stereoscopic conversion of two-dimensional movie encoded in MPEG-2. The stereoscopic algorithm consists of segmentation process and depth determination process. In the segmentation process, we segment the image based on the dc information and the motion vector information encoded by MPEG-2. After the segmentation, depth of each segment is determined by examining the motion vector and the overlapped region of the segment.

[1]  Man-Bae Kim,et al.  Stereoscopic conversion of monoscopic video by the transformation of vertical-to-horizontal disparity , 1998, Electronic Imaging.

[2]  Philip Victor Harman,et al.  Rapid 2D-to-3D conversion , 2002, IS&T/SPIE Electronic Imaging.

[3]  Irfan A. Essa,et al.  Depth layers from occlusions , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[4]  Yi Lu,et al.  Depth detection of targets in a monocular image sequence , 1999, Gateway to the New Millennium. 18th Digital Avionics Systems Conference. Proceedings (Cat. No.99CH37033).

[5]  Yao Wang,et al.  Automatic Video Object Segmentation Using Volume Growing and Hierarchical Clustering , 2004, EURASIP J. Adv. Signal Process..

[6]  Fatih Porikli.  Real-time video object segmentation for MPEG-encoded video sequences , 2004, IS&T/SPIE Electronic Imaging.