Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
xth430 64ab26adee | 1 year ago | |
---|---|---|
opt | 1 year ago | |
svox2 | 1 year ago | |
.gitignore | 1 year ago | |
LICENSE | 1 year ago | |
MANUAL.md | 1 year ago | |
README.md | 1 year ago | |
environment.yml | 1 year ago | |
manual_install.sh | 1 year ago | |
setup.py | 1 year ago | |
teaser.gif | 1 year ago |
Update:
11/10/2022
We added a manual for training, mesh/cage generation and deformation, etc.Project page | Paper | Video
Create a conda environment:
conda env create -f environment.yml
conda activate deforming-nerf
Then install all dependencies (this may take a long time):
pip install .
Note that our code implementation is based on Plenoxels, if you have any installation problems, we recommend referring to their original repo.
Download the dataset from the following links and put them under ./data/
directory:
NeRF-synthetic dataset: https://drive.google.com/drive/folders/128yBriW1IG_3NJ5Rp7APSTZsJqdJdfc1 (nerf_synthetic.zip
)
Synthetic-NSVF dataset: https://github.com/facebookresearch/NSVF#dataset (Synthetic_NSVF.zip
)
DTU dataset (IDR preprocessed): https://github.com/lioryariv/idr (DTU.zip
)
Download optimized Plenoxels models and preprocessed cages from Google Drive and put them under .opt/ckpt/
directory.
Take Lego
in NeRF-synthetic dataset as an example.
Inside opt/
, run the following command.
Render a static deformed scene with moving camera:
python render_imgs_deform.py ckpt/nerf_lego/ckpt.npz ../data/nerf_synthetic/lego/ -c configs/syn.json
Render an animation based on the cage interpolation:
python render_imgs_deform.py ckpt/nerf_lego/ckpt.npz ../data/nerf_synthetic/lego/ -c configs/syn.json --interpolate --cam_id 64
Render the original scene for comparison:
python render_imgs_deform.py ckpt/nerf_lego/ckpt.npz ../data/nerf_synthetic/lego/ -c configs/syn.json --render_orig
Note that you need to specify the config file according to the data format (i.e., syn.json
, syn_nsvf.json
or dtu.json
)
Please see MANUAL.md for some tips about:
If you find our work useful in your research, please consider citing:
@inproceedings{xu2022deforming,
title={Deforming Radiance Fields with Cages},
author={Tianhan Xu and Tatsuya Harada},
year={2022},
booktitle={ECCV},
}
No Description
Text Python Cuda C++ Markdown
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》