Model-based Deep 3D holographic imaging

This is the open source repository for our paper of: "Holographic 3D particle imaging with model-based deep network", Ni Chen, Congli Wang, and Wolfgang Heidrich, IEEE Transactions on Computational Imaging 7: 288-296, 2021.

---

Overview

This repository contains:

• Keras implementation of the MB-HoloNet
• MATLAB implementation of inline hologram synthesize

---

Requirements

• Python 3.6.5
• Tensorflow 1.15
• Keras 2.2.5

How to use the code

• If you have your own data, put it in ./code/data/, then run ./code/keras/train.py after setting your parameters
• If you don't have a code, ./code/matlab/HoloData_sim.m can generate training data.
• You can also down a sample data here: https://drive.google.com/file/d/10i2N8HV9x5AOxZZ5OVOke0LgYENZKZ1E/view?usp=sharing, put it in ./code/data/, and then run the ./code/keras/train.py

---

Contact

Ni Chen, nichen@arizona.edu