Uncertainty quantification of deep learning predictions in photoacoustic imaging

Ruibo Shang; Matthew O'Donnell

doi:10.1117/12.3002000

13 March 2024 Uncertainty quantification of deep learning predictions in photoacoustic imaging

Ruibo Shang, Matthew O'Donnell

Proceedings Volume PC12842, Photons Plus Ultrasound: Imaging and Sensing 2024; PC128422V (2024) https://doi.org/10.1117/12.3002000
Event: SPIE BiOS, 2024, San Francisco, California, United States

Conference Poster

Abstract

Deep learning (DL) is a powerful reconstruction approach now applied broadly in photoacoustic (PA) imaging. However, in practical applications where the ground-truth is unknown, the reliability of predicted PA images from the trained DL network cannot be quantified. Here, we present a new DL approach to simultaneously estimate segmentation (source location) and PA images with uncertainty quantification based on the Bayesian convolutional neural network (BCNN). The BCNN was trained on simulated PA images and tested on both simulated and experimental PA images. The results show accurate segmentation and PA predictions as well as reliable PA uncertainty predictions.

Citation Download Citation

Ruibo Shang and Matthew O'Donnell "Uncertainty quantification of deep learning predictions in photoacoustic imaging", Proc. SPIE PC12842, Photons Plus Ultrasound: Imaging and Sensing 2024, PC128422V (13 March 2024); https://doi.org/10.1117/12.3002000

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