torch-bakedsdf

An unofficial pytorch implementation of Meshing Neural SDFs for Real-Time View Synthesis.

We support exporting baked assets for real-time rendering on WebGL, Unity and Unreal

Install

pip install git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch
pip install -r requirements.txt

For COLMAP, alternative installation options are also available on the COLMAP website

Data preparation

To get COLMAP data from custom images, you should have COLMAP installed (see here for installation instructions). Then put your images in the images/ folder, and run scripts/imgs2poses.py specifying the path containing the images/ folder. For example:

python scripts/imgs2poses.py ./load/bmvs_dog # images are in ./load/bmvs_dog/images

Existing data following this file structure also works as long as images are store in images/ and there is a sparse/ folder for the COLMAP output, for example the data provided by MipNeRF 360.

Run BakedSDF!

python launch.py --config configs/neus-colmap.yaml --gpu 0 --train     dataset.root_dir=$1
python launch.py --config configs/bakedsdf-colmap.yaml --gpu 0 --train     dataset.root_dir=$1 \
                --resume_weights_only --resume latest

Export BakedSDF!

python export.py --exp_dir ./exp/${exp_name}/${trail-name}

for example, when we want to export neus-colmap data, we could run

python export.py --exp_dir ./exp/neus-colmap-stump/@20230907-133647

the export results will be saved in ./results in a glb format

Bring Bakedsdf into your APP!

On Unity and Unreal

You can use BakedSDF2FBX to convert the exported glb and import them into the sample projects of Unity and Unreal

BakedSDF2FBX: http://github.com/AyoubKhammassi/BakedSDF2FBX
UnityBakedSDF: http://github.com/AyoubKhammassi/UnityBakedSDF
UnrealBakedSDF: http://github.com/AyoubKhammassi/UnrealBakedSDF

On Web

The local web viewer is comming soon.

Acknowledgement

The code is based on

@misc{instant-nsr-pl,
    Author = {Yuan-Chen Guo},
    Year = {2022},
    Note = {https://github.com/bennyguo/instant-nsr-pl},
    Title = {Instant Neural Surface Reconstruction}
}

The origin paper:

@article{yariv2023bakedsdf,
  title={BakedSDF: Meshing Neural SDFs for Real-Time View Synthesis},
  author={Yariv, Lior and Hedman, Peter and Reiser, Christian and Verbin, Dor and Srinivasan, Pratul P and Szeliski, Richard and Barron, Jonathan T and Mildenhall, Ben},
  journal={arXiv preprint arXiv:2302.14859},
  year={2023}
}

sanskar107/torch-bakedsdf