Share Email Print

Proceedings Paper

Towards reliable deep learning based phase microscopy (Conference Presentation)

Paper Abstract

We demonstrate a deep-learning(DL)-based computational microscopy for high-throughput phase imaging by taking multiplexed measurements and employing deep neural networks (DNNs) based reconstruction. In particular, we develop a Bayesian convolutional neural network (BNN) to quantify the uncertainties of the DL inference, providing a surrogate estimate of the true prediction errors. The framework is demonstrated on a high-speed computational phase microscopy technique. We show the BNN is able to not only predict high-resolution phase images and but also provide a pixel-wise credibility map that evaluates the imperfections in the datasets and training process。

Paper Details

Date Published: 10 March 2020
Proc. SPIE 11250, High-Speed Biomedical Imaging and Spectroscopy V, 1125015 (10 March 2020); doi: 10.1117/12.2543061
Show Author Affiliations
Yujia Xue, Boston Univ. (United States)
Shiyi Cheng, Boston Univ. (United States)
Yunzhe Li, Boston Univ. (United States)
Lei Tian, Boston Univ. (United States)

Published in SPIE Proceedings Vol. 11250:
High-Speed Biomedical Imaging and Spectroscopy V
Kevin K. Tsia; Keisuke Goda, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?