3D Gaussian Ray Tracing

An implementation of 3D Gaussian Ray Tracing, inspired by the work 3D Gaussian Ray Tracing: Fast Tracing of Particle Scenes. This repository is based on our OptiX-based differentiable 3D Gaussian Ray Tracer

Installation

git clone https://github.com/fudan-zvg/gaussian-raytracing.git --recursive

# This step is same as 3DGS
conda env create --file environment.yml
conda activate gaussian_raytracing

# Install 3DGS's rasterizer
pip install submodules/diff-gaussian-rasterization

# Install 3DGS's simple-knn
pip install submodules/simple-knn

# Install 3D Gaussian Ray Tracer
cd submodules/gtracer && rm -rf ./build && mkdir build && cd build && cmake .. && make && cd ../ && cd ../../
pip install submodules/gtracer

Dataset

Please follow 3DGS for dataset preparation.

Training

# E.g. train a lego
python train.py -s data/nerf_synthetic/lego -m outputs/lego --eval
# E.g. train a lego with GUI
python train.py -s data/nerf_synthetic/lego -m outputs/lego --eval --gui

Evaluating

# Render images
python render.py -m outputs/lego
# Metrics
python metrics.py -m outputs/lego

Interactive Viewer

Use a GUI to view the results.

python gui.py -m outputs/lego

Acknowledgement

• Credits to the original 3D Gaussian Ray Tracing paper.

📜 Citation

If you find this work useful for your research, please cite our github repo:

@misc{gu2024gaussian,
    title = {3D Gaussian Ray Tracing},
    author = {Gu, Chun and Zhang, Li},
    howpublished = {\url{https://github.com/fudan-zvg/gaussian-raytracing}},
    year = {2024}
}