Keynote Speaker

Ezequiel López Rubio

Departamento de Lenguajes y Ciencias de la Computación, Universidad de Málaga

THE RISE OF NEURAL RENDERING IN COMPUTER GRAPHICS

The field of computer graphics has traditionally rooted on classic rendering approaches, where a geometric model of a three-dimensional scene is manually crafted by a digital artist. Such models contain millions of polygons, typically triangles. Then, a classic rendering algorithm produces views of the scene. In contrast to this, neural rendering is a new paradigm for computer graphics that intends to spark a revolution in this workflow. A machine learning model is trained from pictures of a scene, corresponding to different positions and orientations. The model learns the geometry of the scene without human interaction. After that, the model can be queried to generate the desired views. In the past few years, intense research has been done on neural rendering. Many approaches are based on deep learning, while others are founded on advanced geometric and physical frameworks. In this presentation, an overview of the state of the art is provided.