Skip to content

DigitalNatureGroup/CGH-autodiff

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Computer Generated Holograms optimized using Automatic Differentiation (CGH-autodiff)

This repository is a collection of multiple projects that utilize automatic differentiation to optimize Computer Generated Holograms.
There are two projects included (2021/07/23).

Multiple Binary Optimization

"Gradient-based Optimization of Time-Multiplexed Binary Computer-Generated Holograms by Digital Mirror Device"
Kenta Yamamoto, Yoichi Ochiai

This project was orally presented on Digital Holography and Three-Dimensional Imaging at OSA Imaging and Applied Optics Congress.

In this research, we utilize the binary amplitude spatial light modulator (SLM), which has a high refresh rate such as Digital Mirror Device (DMD), to optimize the hologram in time-multiplexing to display a high-definition reproduced image.
The propagation results of multiple binary amplitude holograms are added, and the difference (loss) between the summed image and the target image is taken to optimize each binary amplitude hologram.


Neural Holgoraphy with TensorFlow

This project was developed based on the method of the paper "Neural Holography with Camera-in-the-loop Training" (Paper Link, Source Code, Project Page).
PyTorch was used in "Neural Holography"; however, we implemented it in Tensorflow to use a 2D Fourier Transform that supports automatic differentiation.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published