Intrinsic decomposition and editing of 3D Gaussian splats
Proceedings of the ACM on Computer Graphics and Interactive Techniques (I3D), Volume 9, Number 10 - may 2026
Download the publication :
![paper.pdf [80.5Mo]](/reves/Basilic/images/pdf.png)
Intrinsic decomposition which expresses image colors as the product of diffuse albedo and shading, possibly augmented with
view-dependent residuals has a long history in image editing as it enables the modification of object colors and textures without
altering lighting. We extend intrinsic decomposition to radiance fields represented with Gaussian splatting by proposing solutions to
three key aspects of such decomposition. First, we describe how to model the intrinsic decomposition as independent sets of Gaussian
primitives, which allows each set to adapt to the characteristics of the layer it represents. Second, we present an optimization procedure
guided by data-driven predictions to disentangle multi-view photographs of a scene into the aforementioned intrinsic sets. Finally, we
provide an editing workflow where users modify the texture of planar surfaces simply by modifying the albedo of that surface in one
image. Capturing this edit within the intrinsic radiance field allows re-rendering of the edited scene with plausible lighting under
arbitrary viewpoints.
Images and movies
![Intrinsic3DGS.png [980Ko]](http://www-sop.inria.fr/reves/Basilic/2026/LHLBD26/Intrinsic3DGS.png)
See also
Acknowledgements and Funding
This work was funded by the European Union, European Research Council (ERC) Advanced Grant NERPHYS, 101141721 https://project.inria.fr/nerphys. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. Experiments presented in this paper were carried out using the Grid’5000 testbed, supported by a scientific interest group hosted by Inria and including CNRS, RENATER and several Universities as well as other organizations (see https://www.grid5000.fr). The authors would also like to thank Adobe and NVIDIA for software and hardware donations.
BibTex references
@Article{LHLBD26,
author = "Lanvin, Alexandre and Hu, Jeffrey and Lucas, Simon and Bousseau, Adrien and Drettakis, George",
title = "Intrinsic decomposition and editing of 3D Gaussian splats",
journal = "Proceedings of the ACM on Computer Graphics and Interactive Techniques (I3D)",
number = "10",
volume = "9",
month = "may",
year = "2026",
keywords = "Computing methodologies; Rasterization; Neural networks; Reflectance modeling",
url = "http://www-sop.inria.fr/reves/Basilic/2026/LHLBD26"
}