A Quality-Scalable Depth-Aware Video Processing System

The three-dimensional (3D) displays provide a dramatic improvement of visual quality over the 2D displays. To produce high quality three-dimensional contents, depth map generation is the major issue in content creation. In this paper, we propose a quality-scalable depth-aware video processing system which retrieves depth maps using hybrid depth cues from single view video content. The system uses a hardware-oriented algorithm and generates spectacular and comfortable results. With the help of the depth map, 2D videos can be converted to 3D format and can be displayed in 3D/Stereoscopic displays by depth image-based rendering technique. Moreover, for conventional 2D displays, we also propose a 2D video depth perception enhancement application. With the depth-dependent parameter adjustment, the stereo effect of the 2D content is also enhanced.