A TensorFlow framework for light field deep learning.
lfcnn - A TensorFlow framework for light field deep learning
License and Usage
This software is licensed under the GNU GPLv3 license (see below).
If you use this software in your scientific research, please cite our paper:
Maximilian Schambach and Michael Heizmann: "A Multispectral Light Field Dataset and Framework for Light Field Deep Learning" IEEE Access, vol. 8, pp. 193492-193502, 2020 DOI: 10.1109/ACCESS.2020.3033056
If you use the Normalized Gradient Similarity auxiliary loss training approach in your scientific research, please also cite:
Maximilian Schambach, Jiyang Shi, and Michael Heizmann: "Spectral Reconstruction and Disparity from Spatio-Spectrally Coded Light Fields via Multi-Task Deep Learning" International Conference on 3D Vision (3DV), 2021
Have a look at the Documentation for notes on usage.
Furthermore, you can find some useful examples in the
examples folder which
can help you to get started.
Currently, LFCNN is tested using Python 3.6 -- 3.8 for TensorFlow versions 2.2.0, 2.3.0, 2.4.1, 2.5.1, and 2.6.0, not all of which are all available for Conda (yet).
It likely also work with higher/different versions of Python and TensorFlow.
LFCNN is mostly compatible with all TF versions TensorFlow >= 2.0,
however there is a bug in tf.keras that causes OOMs with data generators
(which LFCNN uses) and multithreading and -processing.
Therefore, we specify
tensorflow >= 2.2 as a dependency,
for which this bug has been resolved.
It is recommended to use Conda to set up a new environment with tensorflow and GPU support. To install with GPU support, run
conda create -n lfcnn python tensorflow-gpu=2.4 tensorflow numpy scipy imageio h5py cudnn cudatoolkit conda activate lfcnn
Then, install LFCNN via
pip install lfcnn
Optionally, for some of LFCNN's features, install the following:
matplotlib(via conda or pip)
pymongo(via conda or pip)
Installation on Windows
TensorFlow under Windows requires the Visual C++ redistributable. Otherwise, follow the general installation notes above.
You can manually run the tests using
$ pytest <path-to-lfcnn>/test/
$ pip uninstall lfcnn
If you are interested in contributing to LFCNN, feel free to create an issue or fork the project and submit a merge request. As this project is still undergoing restructuring and extension, help is always welcome!
Please stick to the PEP 8 Python coding styleguide.
The docstring coding style of the reStructuredText follows the googledoc style.
Copyright (C) 2021 The LFCNN Authors
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.