Continuous Levels of Details for Light Field Networks

Codebase for Continuous Levels of Details for Light Field Networks (BMVC 2022).

Getting Started

1. Download our datasets and extract them to datasets directory.
2. Setup a PyTorch environment and install requirements.txt.
3. To train, run python app.py -c configs/run_continuous_jon.txt.
   Alternatively, download our trained LFNs to runs.

Interactive Viewer

To use the viewer on Ubuntu, run the following:

sudo apt install libmesa-dev libglfw3
# Required to install pycuda with OpenGL support
echo "CUDA_ENABLE_GL = True" > ~/.aksetup-defaults.py
pip install pycuda pyopengl
pip install git+https://github.com/glumpy/glumpy
rm ~/.aksetup-defaults.py

python app.py -c configs/run_continuous_jon.txt --script-mode=viewer

If you get CUBLAS_STATUS_EXECUTION_FAILED while opening the viewer, try running with CUBLAS_WORKSPACE_CONFIG=:0:0. (PyTorch Issue).

Citation

@inproceedings{li2023continuouslodlfn,
author    = {David Li and Brandon Yushan Feng and Amitabh Varshney},
title     = {Continuous Levels of Detail for Light Field Networks},
booktitle = {34th British Machine Vision Conference 2023, {BMVC} 2023, Aberdeen, UK, November 20-24, 2023},
publisher = {{BMVA}},
year      = {2023},
url       = {https://papers.bmvc2023.org/0139.pdf}
}

Acknowledgments

• utils/nerf_utils.py is borrowed from krrish94/nerf-pytorch.