SGCR: Spherical Gaussians for Efficient 3D Curve Reconstruction

Xinran Yang · Donghao ji · Yuanqi Li · Jie Guo · Yanwen Guo · Junyuan Xie

This repository contains the official Pytorch implementation for SGCR: Spherical Gaussians for Efficient 3D Curve Reconstruction (CVPR2025).

Installation

git clone https://github.com/Martinyxr/SGCR.git
cd SGCR
conda env create --file environment.yml 
conda activate SGCR

Demo

# Training Spherical Gaussians 
python train.py -s ./example/00000006 -m ./output/Gaussain/00000006

# 3D Parametric Curve Reconstrcution 
python ./parametric_curve/curve_fitting.py --object_id 00000006

The training command is similar as 3D Gaussian Splatting. After curve reconstruction, the results will be saved in ./output/curve/.

Citation

if you find the code useful, please consider the following BibTeX entry.

@InProceedings{yang2025sgcr,
  title        = {SGCR: Spherical Gaussians for Efficient 3D Curve Reconstruction},
  author       = {Yang, Xinran and Ji, Donghao and Li, Yuanqi and Guo, Jie and Guo, Yanwen and Xie, Junyuan},
  booktitle    = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year         = {2025}
}

Acknowledgments

This project is built upon 3DGS. The evaluation ABC-NEF dataset is from NEF. We use pretrained PidiNet for edge map extraction. We thank all the authors for their great work and repos.