RELIGHTABLE NEURAL RADIANCE FIELD MODEL

The present invention sets forth a technique for generating two-dimensional (2D) renderings of a three-dimensional (3D) scene from an arbitrary camera position under arbitrary lighting conditions. This technique includes determining, based on a plurality of 2D representations of a 3D scene, a radian...

Full description

Saved in:
Bibliographic Details
Main Authors ZOSS, Gaspard, URNAU GOTARDO, Paulo Fabiano, CHANDRAN, Prashanth, BRADLEY, Derek Edward, XU, Yingyan
Format Patent
LanguageEnglish
Published 16.05.2024
Subjects
CALCULATING
COMPUTING
COUNTING
IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The present invention sets forth a technique for generating two-dimensional (2D) renderings of a three-dimensional (3D) scene from an arbitrary camera position under arbitrary lighting conditions. This technique includes determining, based on a plurality of 2D representations of a 3D scene, a radiance field function for a neural radiance field (NeRF) model. This technique further includes determining, based on a plurality of 2D representations of a 3D scene, a radiance field function for a "one light at a time" (OLAT) model. The technique further includes rendering a 2D representation of the scene based on a given camera position and illumination data. The technique further includes computing a rendering loss based on the difference between the rendered 2D representation and an associated one of the plurality of 2D representations of the scene. The technique further includes modifying at least one of the NeRF and OLAT models based on the rendering loss.
Bibliography:Application Number: US202318505009