3D Visual Communications

2.1 3DTV Content Processing Procedure

Stereoscopic 3DTV has to provide each eye with a separate 2D video stream for the perception of the 3D world. The easiest way to represent the 3D content is to prepare the whole set of videos for all possible views. However, this representation requires high bandwidth and may not be suitable for interactively viewing the scene. Additionally, naked-eye free-view 3DTV will appear in the market soon and this kind of equipment will require multiple views for creating the sense of seeing the 3D world from different viewpoints. Thus, we must first understand the pipeline of 3DTV processing procedure which involves four basic tasks: representation, transmission, animation, and rendering as shown in Figure 2.2.

1. Representation

   A real world scene contains moving objects with different shapes and complexities. Generally, the best choice of representation scheme depends on the requirement of the applications including the properties of the scene, such as the complexity, appearance and motion the transmission speed, the requirement of realism, and interactivity. A typical example is to show a view of realistic scenery to an end user such as a view of the Grand Canyon. Since the requirement of the application is realistic, the scene is static and comprised geometries are generally far away, a set of captured images, that is, image-based representations, are probably the best choice. However, in a 3D online urban shooting game, proper interactivity is the ...