Friday, October 7, 2022
HomeTechSuperior phase recovery and hologram reconstruction using a deep neural network

Superior phase recovery and hologram reconstruction using a deep neural network


Fourier Imager Community (FIN): A deep neural community for hologram reconstruction with superior exterior generalization. Credit: Ozcan Lab @ UCLA

Deep studying has achieved benchmark outcomes for varied imaging duties, together with holographic microscopy, the place a vital step is to get well the part info of samples utilizing intensity-only measurements. By coaching on well-designed datasets, deep neural networks have confirmed to outperform classical part retrieval and hologram reconstruction algorithms when it comes to accuracy and computational effectivity. Nevertheless, mannequin generalization, which refers to extending the neural networks’ capabilities to new forms of samples by no means seen throughout the coaching, stays a problem for current deep studying fashions.

UCLA researchers have lately created a novel neural network structure, termed Fourier Imager Community (FIN), which demonstrated unprecedented generalization to unseen pattern sorts, additionally reaching superior computational pace in part retrieval and holographic picture reconstruction duties. On this new strategy, they launched spatial Fourier rework modules that allow the neural community to reap the benefits of the spatial frequencies of the entire picture. UCLA researchers skilled their FIN mannequin on human lung tissue samples and demonstrated its superior generalization by reconstructing the holograms of human prostate and salivary gland tissue sections, and Pap smear samples, which have been by no means seen within the coaching part.

Revealed in Mild: Science & Purposes, this new deep learning-based framework is reported to attain increased picture reconstruction accuracy in comparison with the classical hologram reconstruction algorithms and the state-of-the-art deep learning models, whereas shortening the reconstruction time by ~50 occasions. This new deep studying framework will be broadly used to create extremely generalizable neural networks for varied microscopic imaging and pc imaginative and prescient duties.

This analysis was led by Dr. Aydogan Ozcan, Chancellor’s Professor and Volgenau Chair for Engineering Innovation at UCLA and HHMI Professor with the Howard Hughes Medical Institute. The opposite authors of this work embrace Hanlong Chen, Luzhe Huang, and Tairan Liu, all from the Electrical and Pc Engineering division at UCLA. Prof. Ozcan additionally has UCLA school appointments within the bioengineering and surgical procedure departments and is an affiliate director of the California NanoSystems Institute.


Faster holographic imaging using recurrent neural networks


Extra info:
Hanlong Chen et al, Fourier Imager Community (FIN): A deep neural community for hologram reconstruction with superior exterior generalization, Mild: Science & Purposes (2022). DOI: 10.1038/s41377-022-00949-8

Quotation:
Superior part restoration and hologram reconstruction utilizing a deep neural community (2022, August 16)
retrieved 16 August 2022
from https://techxplore.com/information/2022-08-superior-phase-recovery-hologram-reconstruction.html

This doc is topic to copyright. Other than any truthful dealing for the aim of personal examine or analysis, no
half could also be reproduced with out the written permission. The content material is offered for info functions solely.



Click Here To Join Our Telegram Channel



Source link

When you have any considerations or complaints concerning this text, please tell us and the article might be eliminated quickly. 

Raise A Concern

RELATED ARTICLES
- Advertisment -

Most Popular