Research Area:  Machine Learning

Abstract:

In this article we present a survey of the different techniques of rendering of 3D computer generated images. We start with the principles and advances of the traditional methods of rasterization and ray tracing. Then, we discover the new techniques based on deep learning, which are now part of a new discipline of computer graphics called neural rendering, allowing the synthesis and rendering of 3D images, thanks to generative adversarial network and variational auto encoder models. Finally, we compare theses approaches according to different criteria.

Keywords:  

Author(s) Name:  Houssam Halmaoui & Abdelkrim Haqiq

Publisher name:  Springer

DOI:  10.1007/978-3-030-96299-9_51

Volume Information:  

Paper Link:   https://link.springer.com/chapter/10.1007/978-3-030-96299-9_51