A comparison of prediction structures for multi-view video coding based on the H.264/AVC standard

Abstract

The deployment of 3D techniques is one of the most promising fields regarding the development of new applications for natural video scenes, as they provide users greatly enhanced viewing experience. For such applications, multi-view video coding is crucial to handle the enormous amount of data. This paper provides an experimental analysis of multi-view video coding for various temporal and interview prediction structures. We present the advantage of several multi-view video coding prediction schemes. This demonstration will show the benefits of using hierarchical B pictures based on H.264/AVC.