NeuS

We present a novel neural surface reconstruction method, called NeuS (pronunciation: /nuːz/, same as "news"), for reconstructing objects and scenes with high fidelity from 2D image inputs.

Project page | Paper | Data

This is the official repo for the implementation of NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction.

Usage

Data Convention

The data is organized as follows:

<case_name>
|-- cameras_xxx.npz    # camera parameters
|-- image
    |-- 000.png        # target image for each view
    |-- 001.png
    ...
|-- mask
    |-- 000.png        # target mask each view (For unmasked setting, set all pixels as 255)
    |-- 001.png
    ...

Here the cameras_xxx.npz follows the data format in IDR, where world_mat_xx denotes the world to image projection matrix, and scale_mat_xx denotes the normalization matrix.

Setup

Clone this repository

git clone https://github.com/Totoro97/NeuS.git
cd NeuS
pip install -r requirements.txt

Dependencies (click to expand)

• torch==1.8.0
• opencv_python==4.5.2.52
• trimesh==3.9.8
• numpy==1.19.2
• pyhocon==0.3.57
• icecream==2.1.0
• tqdm==4.50.2
• scipy==1.7.0
• PyMCubes==0.1.2

Running

• Training without mask

python exp_runner.py --mode train --conf ./confs/womask.conf --case <case_name>

• Training with mask

python exp_runner.py --mode train --conf ./confs/wmask.conf --case <case_name>

• Extract surface from trained model

python exp_runner.py --mode validate_mesh --conf <config_file> --case <case_name> --is_continue # use latest checkpoint

The corresponding mesh can be found in exp/<case_name>/<exp_name>/meshes/<iter_steps>.ply.

• View interpolation

python exp_runner.py --mode interpolate_<img_idx_0>_<img_idx_1> --conf <config_file> --case <case_name> --is_continue # use latest checkpoint

The corresponding image set of view interpolation can be found in exp/<case_name>/<exp_name>/render/.

Train NeuS with your custom data

More information can be found in preprocess_custom_data.

Citation

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

@article{wang2021neus,
  title={NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction},
  author={Wang, Peng and Liu, Lingjie and Liu, Yuan and Theobalt, Christian and Komura, Taku and Wang, Wenping},
  journal={arXiv preprint arXiv:2106.10689},
  year={2021}
}

Acknowledgement

Some code snippets are borrowed from IDR and NeRF-pytorch. Thanks for these great projects.