Site en français


LISA - Virtual Reality

LISA (Laboratory of Image Synthesis and Analysis) brings together expertise in image processing and analysis, pattern recognition, image synthesis and virtual reality. 
LISA-VR studies advanced Computational Imaging techniques to capture and render natural scenes in 3D for Virtual Reality (cf. attached pictures 1 & 2) and Holography (cf. attached pictures 3 & 4). Instead of explicitly modeling the scene with 3D models as is typically done in 3D video games, we aim at low-cost light-field solutions, characterizing not only the light intensity, but also the light direction in each point in space, aka the plenoptic function. To obtain a specific view to the scene, only a subset of the various light directions is rendered, using so-called Depth Image-Based Rendering techniques that only use images to create proper 3D perspective illusions. Their principles are know for a long time, but it remains challenging to make them work in practical settings; in particular the depth sensing/estimation, the disocclusion handling and the camera architectural settings remain open research questions. The ultimate goal is to be able to properly capture the 3D scene's light field with as little equipment as possible, while reaching high-resolution virtual views of the scene for Holographic portraiting and Virtual Reality, the latter also imposing stringent real-time rendering constraints.


This person isn't currently part of a projet.