Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

3iGS: Factorised Tensorial Illumination for 3D Gaussian Splatting

About

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.

Zhe Jun Tang, Tat-Jen Cham• 2024

Related benchmarks

TaskDatasetResultRank
Novel View SynthesisNeRF-synthetic original (test)
PSNR33.6
25
Reflective Object ReconstructionGlossy Synthetic
PSNR26.86
15
Novel View SynthesisShiny Blender
PSNR30.64
13
Modeling reflectionsRef-NeRF
PSNR30.91
9
Novel View SynthesisShiny Real
PSNR24.1
8
Modeling reflectionsRef-Real
PSNR23.67
8
Novel View SynthesisGlossy Synthetic original (test)
PSNR26.39
8
Normal ReconstructionShiny Blender (ShinyB) original (test)
MAE15.47
7
Radiance-field-based reconstructionReflective Benchmarks (test)
Rendering Speed0.44
6
Novel View SynthesisSynthetic datasets (test)
Rendering Speed (x)0.44
5
Showing 10 of 10 rows

Other info

Follow for update