SIGGRAPH 2024
# build environment with python 3.7
conda create -n pinf python=3.7
conda activate pinf
# if ffmpeg is not installed (test by ffmpeg -version)
conda install -c conda-forge ffmpeg
conda install ffmpeg
# requirments
pip install -r requirments
# raymarching
cd raymarching
pip install -e .
# test environment
python env_test.py
The dataset used in the paper can be downloaded from Goolge Drive.
Take the Cylinder scene as an example:
python train.py --config configs/cyl.txt
# velocity voxel output
python test.py --config configs/cyl.txt --testskip 1 --output_voxel --full_vol_output
# render novel view
python test.py --config configs/cyl.txt --render_only
# static object mesh
python test.py --config configs/cyl.txt --mesh_only
- Ninja is required to load C++ extensions
pip install Ninja
@inproceedings{Wang2024PICT,
author = {Wang, Yiming and Tang, Siyu and Chu, Mengyu},
title = {Physics-Informed Learning of Characteristic Trajectories for Smoke Reconstruction},
year = {2024},
url = {https://doi.org/10.1145/3641519.3657483},
doi = {10.1145/3641519.3657483},
booktitle = {ACM SIGGRAPH 2024 Conference Papers},
articleno = {53},
numpages = {11},
series = {SIGGRAPH '24}
}