Sample and system parameter estimation from local speckle pattern by fully numerically trained deep convolution neural network

T. Seesan; D. Oida; K. Oikawa; P. Buranasiri; Y. Yasuno

doi:10.1117/12.2578012

5 March 2021 Sample and system parameter estimation from local speckle pattern by fully numerically trained deep convolution neural network

T. Seesan, D. Oida, K. Oikawa, P. Buranasiri, Y. Yasuno

Author Affiliations +

Proceedings Volume 11630, Optical Coherence Tomography and Coherence Domain Optical Methods in Biomedicine XXV; 1163021 (2021) https://doi.org/10.1117/12.2578012
Event: SPIE BiOS, 2021, Online Only

Conference Poster

Abstract

Deep convolutional neural network (CNN) based estimators for optical coherence tomography (OCT) are presented. The CNN analyze local OCT speckle patterns and estimate the sample’s scatterer density and OCT resolutions. This estimator is intensity invariant, i.e., it does not use the net signal strength of OCT even to estimate the scatterer density. The CNN is trained by a huge training dataset that was generated by a simple simulator of OCT imaging. And hence, the CNN is trained without experimental datasets. The performance of CNN was evaluated by numerically generated OCT images, and good accuracies of the estimation were shown.

Conference Presentation

Citation Download Citation

T. Seesan, D. Oida, K. Oikawa, P. Buranasiri, and Y. Yasuno "Sample and system parameter estimation from local speckle pattern by fully numerically trained deep convolution neural network", Proc. SPIE 11630, Optical Coherence Tomography and Coherence Domain Optical Methods in Biomedicine XXV, 1163021 (5 March 2021); https://doi.org/10.1117/12.2578012

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