SIGGRAPH 2026

Project Page | Paper | arXiv | Code

We present a novel differentiable framework to automatically learn view-dependent 2D kernels in a splatting-based pipeline, that are tailored to efficiently represent various types of scenes.

Branches

This repository organizes code by experiment type using separate branches. Each branch is independent and can be cloned directly:

git clone -b <branch-name> https://github.com/optkernel/codebase

Branch	Description
2d_splatting	2D splatting experiments
3d_splatting	3D splatting experiments
2d_image_representation	2D image representation experiments

Please refer to the README in each branch for environment setup and usage instructions.

Citation

Cite as below if you find this repository is helpful to your project:

@inproceedings{ding2026kernel,
    title     = {Learning View-Dependent Splatting Kernels},
    author    = {Huakeng Ding and Zhangpeng Liu and Fan Pei and Kun Zhou and Hongzhi Wu},
    booktitle = {SIGGRAPH 2026 Conference Papers},
    year      = {2026}
}

Acknowledgments

We have intensively borrowed code from 3D gaussian splatting, 2D Gaussian Splatting, Beta-Splatting and 3DGS-MCMC. Many thanks to the authors for sharing their codes.