Upload license.

LICENSE

@@ -0,0 +1,21 @@
+Permission is hereby granted, free of charge, to any person or organization
+obtaining a copy of the software and accompanying documentation covered by
+this license (the "Software") to use, reproduce, display, distribute,
+execute, and transmit the Software, and to prepare derivative works of the
+Software, and to permit third-parties to whom the Software is furnished to
+do so, all subject to the following:
+
+The copyright notices in the Software and this entire statement, including
+the above license grant, this restriction and the following disclaimer,
+must be included in all copies of the Software, in whole or in part, and
+all derivative works of the Software, unless such copies or derivative
+works are solely in the form of machine-executable object code generated by
+a source language processor.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT
+SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE
+FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+DEALINGS IN THE SOFTWARE.

README.md

@@ -5,16 +5,15 @@
 
 This code implements a differentiable proxy model for simulating meta-optics and a neural feature propagation deconvolution method. These components are optimized end-to-end using machine learning optimizers.
 
-The experimental results from the manuscript and the supplemental information are reproducible with this implementation. The proposed differentiable proxy model, neural feature propagation, and end-to-end optimization framework are implemented completely in TensorFlow, without dependency on third-party libraries. While the fastest inference and training times are achieved by running on state-of-the-art GPUs, this code can be run solely with standard TensorFlow packages, ensuring high portability and ease of reproducibility.
-
+The experimental results from the manuscript and the supplemental information are reproducible with this implementation. The proposed differentiable proxy model, neural feature propagation, and end-to-end optimization framework are implemented completely in TensorFlow, without dependency on third-party libraries.
 
 ## Training
 To perform end-to-end training (of meta-optic and deconvolution) execute the 'run_train.sh' script. The model checkpoint which includes saved parameters for both the meta-optic and deconvolution will be saved to 'training/ckpt'. The folder 'training/data' contains a subset of the training and test data that we used for optimizing our end-to-end imaging pipeline.
 
-
 ## Testing
 To perform inference on real-world captures launch the "test.ipynb" notebook in Jupyter Notebook and step through the cells. The notebook will load in a finetuned checkpoint of our neural feature propagation network from 'experimental/ckpt' which will process captured sensor measurements located in 'experimental/data'. The reconstructed images will be displayed within the notebook.
 
+Additional captured sensor measurements can be found in the [data repository](https://drive.google.com/drive/folders/1fsAvN9MPtN5jJPeIFjWuLUY9Hp8NNkar?usp=sharing).
 
 ## Requirements
 This code has been tested with Python 3.6.10 using TensorFlow 2.2.0 running on Linux with an Nvidia P100 GPU with 16GB RAM.
@@ -29,3 +28,21 @@ Scipy
 matplotlib
 jupyter-notebook
 ```
+
+## Citation
+If you find our work useful in your research, please cite:
+```
+@article{Tseng2021NeuralNanoOptics,
+    title   = "Neural Nano-Optics for High-quality Thin Lens Imaging",
+    author  = "Tseng, Ethan and Colburn, Shane and Whitehead, James and Huang, Luocheng
+               and Baek, Seung-Hwan and Majumdar, Arka and Heide, Felix",
+    journal = "Nature Communications",
+    volume  = ,
+    number  = ,
+    pages   = ,
+    year    = 2021
+}
+```
+
+## License
+Our code is licensed under BSL-1. By downloading the software, you agree to the terms of this License. The training data in the folder 'training/data' comes from the [INRIA Holidays Dataset](https://lear.inrialpes.fr/~jegou/data.php).

test.ipynb

@@ -6,7 +6,7 @@
    "source": [
     "## Inference Code for 'Neural Nano-Optics for High-quality Thin Lens Imaging'\n",
     "\n",
-    "#### This Jupyter Notebook can be used to produce the experimental reconstructions shown in the manuscript and in the supplemental information."
+    "#### This notebook can be used to produce the experimental reconstructions shown in the manuscript and in the supplemental information."
    ]
   },
   {