3D Images Compression for Multi-View Auto-Stereoscopic Displays

As auto-stereoscopic displays offer more and more different possible views, the problem of the size of the data to display is becoming more and more urging. If many 3D images compression algorithms already exist, using various methods, many are not lossless and lead to zones where data are not well displayed. This article details our lossless 3D compression algorithm based on 3D reconstruction of the initial 3D scene and presents some results. We will insist on the description of the coding and decoding process of a number of particular effects, especially the reflection and refraction reconstitution. We will also show an encountered problem called discontinuity of the facets and the way to solve it