Code for 'Embedding Deep Learning in Inverse Scattering Problem'.

1. Link for original article - https://ieeexplore.ieee.org/document/8709721
2. Slides for conference presentation - https://sanghviyashiitb.github.io/blog/2019-3-31-URSI

Instructions

1. Before running the code, ensure that Python-3.5+, Jupyter Notebook is installed along with the necessary packages i.e.
   • numpy
   • scipy
   • matplotlib
   • pytorch
   • PIL
2. Download the repository into your local system as zip file and unpack it. OR clone the git reporsitory using the following command:

git clone https://github.com/sanghviyashiitb/EmbeddingDLinISP-Github.git

3. Enter the directory as

cd EmbeddingDLinISP-Github/

4. Run file download_model.py to download the trained CS-Net.

python3 download_model.py

4. Open Tutorial.ipynb as a jupyter notebook to use the code provided!

jupyter notebook

The python script for downloading model file was provided by user turdus-merula from the link here.

License

The code provided in this repository (i.e. Python and Jupyter scripts) is released under the MIT License

Citation

If you're using the inverse scattering code, please cite us as follows:
Journal Article
@article{sanghvi2019embedding,
title={Embedding Deep Learning in Inverse Scattering Problems},
author={Sanghvi, Yash and Kalepu, Yaswanth N Ganga Bhavani and Khankhoje, Uday},
journal={IEEE Transactions on Computational Imaging},
year={2019},
publisher={IEEE}
}

Also feel free to send your questions/feedback about the code or the paper to [email protected] !