Title	SuGaR
Author	`Gu{\'e}don, Antoine and Lepetit, Vincent`
Conf/Jour	CVPR
Year	2024
Project	Anttwo/SuGaR: Official implementation of SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering (github.com)
Paper	SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering

实验(win10)

~~conda env create -f environment.yml~~
复制environment.yml中pip包到requirements.txt中，通过pip安装
conda create -n sugar
pip install -r requirements.txt
add:

plyfile
tqdm
rich

pip install torch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118

pytorch3d: pytorch3D Windows下安装经验总结_windows安装pytorch3d-CSDN博客

cd gaussian_splatting/submodules/diff-gaussian-rasterization/
pip install -e .
cd ../simple-knn/
pip install -e .
cd ../../../

run

# GS
python gaussian_splatting/train.py -s <path to COLMAP or NeRF Synthetic dataset> --iterations 7000 -m <path to the desired output directory>

-

## eg
python gaussian_splatting/train.py -s inputs/dtu_scan114 --iterations 7000 -m exp/dtu_scan114
python gaussian_splatting/train.py -s inputs/Miku --iterations 7000 -m exp/Miku

# SuGaR
python train.py -s <path to COLMAP or NeRF Synthetic dataset> -c <path to the Gaussian Splatting checkpoint> -r <"density" or "sdf">

##eg
python train.py -s inputs/dtu_scan114 -c exp/dtu_scan114/ -r sdf
python train.py -s inputs/Miku -c exp/Miku/ -r sdf