Show simple item record

dc.contributor.authorRodriguez-Pardo, Carlosen_US
dc.contributor.authorFabre, Javieren_US
dc.contributor.authorGarces, Elenaen_US
dc.contributor.authorLopez-Moreno, Jorgeen_US
dc.contributor.editorRitschel, Tobiasen_US
dc.contributor.editorWeidlich, Andreaen_US
dc.date.accessioned2023-06-27T07:03:01Z
dc.date.available2023-06-27T07:03:01Z
dc.date.issued2023
dc.identifier.issn1467-8659
dc.identifier.urihttps://doi.org/10.1111/cgf.14883
dc.identifier.urihttps://diglib.eg.org:443/handle/10.1111/cgf14883
dc.description.abstractEnvironment maps are commonly used to represent and compute far-field illumination in virtual scenes. However, they are expensive to evaluate and sample from, limiting their applicability to real-time rendering. Previous methods have focused on compression through spherical-domain approximations, or on learning priors for natural, day-light illumination. These hinder both accuracy and generality, and do not provide the probability information required for importance-sampling Monte Carlo integration. We propose NEnv, a deep-learning fully-differentiable method, capable of compressing and learning to sample from a single environment map. NEnv is composed of two different neural networks: A normalizing flow, able to map samples from uniform distributions to the probability density of the illumination, also providing their corresponding probabilities; and an implicit neural representation which compresses the environment map into an efficient differentiable function. The computation time of environment samples with NEnv is two orders of magnitude less than with traditional methods. NEnv makes no assumptions regarding the content (i.e. natural illumination), thus achieving higher generality than previous learning-based approaches. We share our implementation and a diverse dataset of trained neural environment maps, which can be easily integrated into existing rendering engines.en_US
dc.publisherThe Eurographics Association and John Wiley & Sons Ltd.en_US
dc.subjectCCS Concepts: Computing methodologies -> Neural networks; Image-based rendering; Image representations
dc.subjectComputing methodologies
dc.subjectNeural networks
dc.subjectImage
dc.subjectbased rendering
dc.subjectImage representations
dc.titleNEnv: Neural Environment Maps for Global Illuminationen_US
dc.description.seriesinformationComputer Graphics Forum
dc.description.sectionheadersNeural Rendering
dc.description.volume42
dc.description.number4
dc.identifier.doi10.1111/cgf.14883
dc.identifier.pages14 pages


Files in this item

Thumbnail

This item appears in the following Collection(s)

  • 42-Issue 4
    Rendering 2023 - Symposium Proceedings

Show simple item record