Light-field video coding using geometry-based disparity compensation

Abstract

One of the main challenges in 3D light-field imaging approaches lies in the massive amount of visual information involved in providing 3D content with sufficient resolution. Consequently, adequate coding tools are essential for efficient transmission and storage of this type of content. In this context, this paper presents and evaluates two coding solutions based on the High Efficiency Video Coding (HEVC) scheme and for efficient compression of the 3D light-field content. These two coding schemes aim to exploit the 3D geometry-based disparity information in the 3D light field content and replace the block-based disparity estimation. In the first scheme, the disparity map of each view is used to directly derive the vectors for compensation, and in the second scheme these disparity vectors (for all views) are calculated (for non-occluded areas) from the disparity map of the base view. A comparative study of these proposed coding schemes is performed and future research directions are also discussed.