Computer generated hologram compression with attention-based deep convolutional neural network

Zhelun Shen; Guanglin Yang; Haiyan Xie

doi:10.1117/12.2602659

9 October 2021 Computer generated hologram compression with attention-based deep convolutional neural network

Zhelun Shen, Guanglin Yang, Haiyan Xie

Proceedings Volume 11898, Holography, Diffractive Optics, and Applications XI; 118981Y (2021) https://doi.org/10.1117/12.2602659
Event: SPIE/COS Photonics Asia, 2021, Nantong, JS, China

Conference Poster

Abstract

We propose an attention-based deep convolutional neural network for computer generated hologram (CGH) compression, where a channel attention mechanism is applied to both computer generated hologram compression and reconstruction. By applying deep convolutional neural networks in the compression process, we can extract more compact and representative information than bicubic interpolation. Additionally, a channel attention mechanism is applied to selectively emphasize informative features and suppress less useful ones for both CGH compression and reconstruction. By employing attention mechanisms to enhance the feature representation ability of deep convolutional neural networks, we can further improve the performance of the reconstructed computer generated hologram. Experimental results show our method can better recover the compressed computer generated hologram than only employing convolutional neural networks in the reconstruction process.

Citation Download Citation

Zhelun Shen, Guanglin Yang, and Haiyan Xie "Computer generated hologram compression with attention-based deep convolutional neural network", Proc. SPIE 11898, Holography, Diffractive Optics, and Applications XI, 118981Y (9 October 2021); https://doi.org/10.1117/12.2602659

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available