3iGS: factorised tensorial illumination for 3D Gaussian splatting

The use of 3D Gaussians as representation of radiance fields has enabled high quality novel view synthesis at real-time rendering speed. However, the choice of optimising the outgoing radiance of each Gaussian independently as spherical harmonics results in unsatisfactory view dependent effects. In...

Full description

Saved in:
Bibliographic Details
Main Authors: Tang, Zhe Jun, Cham, Tat-Jen
Other Authors: College of Computing and Data Science
Format: Conference or Workshop Item
Language:English
Published: 2025
Subjects:
Online Access:https://hdl.handle.net/10356/182915
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-182915
record_format dspace
spelling sg-ntu-dr.10356-1829152025-03-09T03:24:37Z 3iGS: factorised tensorial illumination for 3D Gaussian splatting Tang, Zhe Jun Cham, Tat-Jen College of Computing and Data Science 18th European Conference on Computer Vision (ECCV 2024) S-Lab Computer and Information Science 3D Gaussians Computer graphics The use of 3D Gaussians as representation of radiance fields has enabled high quality novel view synthesis at real-time rendering speed. However, the choice of optimising the outgoing radiance of each Gaussian independently as spherical harmonics results in unsatisfactory view dependent effects. In response to these limitations, our work, Factorised Tensorial Illumination for 3D Gaussian Splatting, or 3iGS, improves upon 3D Gaussian Splatting (3DGS) rendering quality. Instead of optimising a single outgoing radiance parameter, 3iGS enhances 3DGS view-dependent effects by expressing the outgoing radiance as a function of a local illumination field and Bidirectional Reflectance Distribution Function (BRDF) features. We optimise a continuous incident illumination field through a Tensorial Factorisation representation, while separately fine-tuning the BRDF features of each 3D Gaussian relative to this illumination field. Our methodology significantly enhances the rendering quality of specular view-dependent effects of 3DGS, while maintaining rapid training and rendering speeds. Nanyang Technological University This study is supported under the RIE2020 Industry Alignment Fund - Industry Collaboration Projects (IAF-ICP) Funding initiative, as well as cash and in-kind collaboration from the industry partner(s). 2025-03-09T03:24:37Z 2025-03-09T03:24:37Z 2025 Conference Paper Tang, Z. J. & Cham, T. (2025). 3iGS: factorised tensorial illumination for 3D Gaussian splatting. 18th European Conference on Computer Vision (ECCV 2024), LNCS 15072, 143-159. https://dx.doi.org/10.1007/978-3-031-72630-9_9 9783031726293 https://hdl.handle.net/10356/182915 10.1007/978-3-031-72630-9_9 2-s2.0-85212978530 LNCS 15072 143 159 en © 2025 The Author(s), under exclusive license to Springer Nature Switzerland AG. All rights reserved.
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Computer and Information Science
3D Gaussians
Computer graphics
spellingShingle Computer and Information Science
3D Gaussians
Computer graphics
Tang, Zhe Jun
Cham, Tat-Jen
3iGS: factorised tensorial illumination for 3D Gaussian splatting
description The use of 3D Gaussians as representation of radiance fields has enabled high quality novel view synthesis at real-time rendering speed. However, the choice of optimising the outgoing radiance of each Gaussian independently as spherical harmonics results in unsatisfactory view dependent effects. In response to these limitations, our work, Factorised Tensorial Illumination for 3D Gaussian Splatting, or 3iGS, improves upon 3D Gaussian Splatting (3DGS) rendering quality. Instead of optimising a single outgoing radiance parameter, 3iGS enhances 3DGS view-dependent effects by expressing the outgoing radiance as a function of a local illumination field and Bidirectional Reflectance Distribution Function (BRDF) features. We optimise a continuous incident illumination field through a Tensorial Factorisation representation, while separately fine-tuning the BRDF features of each 3D Gaussian relative to this illumination field. Our methodology significantly enhances the rendering quality of specular view-dependent effects of 3DGS, while maintaining rapid training and rendering speeds.
author2 College of Computing and Data Science
author_facet College of Computing and Data Science
Tang, Zhe Jun
Cham, Tat-Jen
format Conference or Workshop Item
author Tang, Zhe Jun
Cham, Tat-Jen
author_sort Tang, Zhe Jun
title 3iGS: factorised tensorial illumination for 3D Gaussian splatting
title_short 3iGS: factorised tensorial illumination for 3D Gaussian splatting
title_full 3iGS: factorised tensorial illumination for 3D Gaussian splatting
title_fullStr 3iGS: factorised tensorial illumination for 3D Gaussian splatting
title_full_unstemmed 3iGS: factorised tensorial illumination for 3D Gaussian splatting
title_sort 3igs: factorised tensorial illumination for 3d gaussian splatting
publishDate 2025
url https://hdl.handle.net/10356/182915
_version_ 1826362242830958592