Enabling fast and high quality LED photoacoustic imaging: A recurrent neural networks based approach

Emran Mohammad Abu Anas, Haichong K. Zhang, Jin Kang, Emad Boctor

Research output: Contribution to journalArticle

Abstract

Photoacoustic (PA) techniques have shown promise in the imaging of tissue chro-mophores and exogenous contrast agents in various clinical applications. However, the key drawback of current PA technology is its dependence on a complex and hazardous laser system for the excitation of a tissue sample. Although light-emitting diodes (LED) have the potential to replace the laser, the image quality of an LED-based system is severely corrupted due to the low output power of LED elements. The current standard way to improve the quality is to increase the scanning time, which leads to a reduction in the imaging speed and makes the images prone to motion artifacts. To address the challenges of longer scanning time and poor image quality, in this work we present a deep neural networks based approach that exploits the temporal information in PA images using a recurrent neural network. We train our network using 32 phantom experiments; on the test set of 30 phantom experiments, we achieve a gain in the frame rate of 8 times with a mean peak-signal-to-noise-ratio of 35.4 dB compared to the standard technique.

Original languageEnglish (US)
Article number331886
Pages (from-to)3852-3866
Number of pages15
JournalBiomedical Optics Express
Volume9
Issue number8
DOIs
StatePublished - Aug 1 2018

Fingerprint

Photoacoustic Techniques
light emitting diodes
Light
Lasers
Signal-To-Noise Ratio
Artifacts
Contrast Media
scanning
lasers
artifacts
signal to noise ratios
output
excitation

ASJC Scopus subject areas

  • Biotechnology
  • Atomic and Molecular Physics, and Optics

Cite this

Enabling fast and high quality LED photoacoustic imaging : A recurrent neural networks based approach. / Anas, Emran Mohammad Abu; Zhang, Haichong K.; Kang, Jin; Boctor, Emad.

In: Biomedical Optics Express, Vol. 9, No. 8, 331886, 01.08.2018, p. 3852-3866.

Research output: Contribution to journalArticle

Anas, Emran Mohammad Abu ; Zhang, Haichong K. ; Kang, Jin ; Boctor, Emad. / Enabling fast and high quality LED photoacoustic imaging : A recurrent neural networks based approach. In: Biomedical Optics Express. 2018 ; Vol. 9, No. 8. pp. 3852-3866.
@article{75dea35c0cd7486cb8fb026772e8bc62,
title = "Enabling fast and high quality LED photoacoustic imaging: A recurrent neural networks based approach",
abstract = "Photoacoustic (PA) techniques have shown promise in the imaging of tissue chro-mophores and exogenous contrast agents in various clinical applications. However, the key drawback of current PA technology is its dependence on a complex and hazardous laser system for the excitation of a tissue sample. Although light-emitting diodes (LED) have the potential to replace the laser, the image quality of an LED-based system is severely corrupted due to the low output power of LED elements. The current standard way to improve the quality is to increase the scanning time, which leads to a reduction in the imaging speed and makes the images prone to motion artifacts. To address the challenges of longer scanning time and poor image quality, in this work we present a deep neural networks based approach that exploits the temporal information in PA images using a recurrent neural network. We train our network using 32 phantom experiments; on the test set of 30 phantom experiments, we achieve a gain in the frame rate of 8 times with a mean peak-signal-to-noise-ratio of 35.4 dB compared to the standard technique.",
author = "Anas, {Emran Mohammad Abu} and Zhang, {Haichong K.} and Jin Kang and Emad Boctor",
year = "2018",
month = "8",
day = "1",
doi = "10.1364/BOE.9.003852",
language = "English (US)",
volume = "9",
pages = "3852--3866",
journal = "Biomedical Optics Express",
issn = "2156-7085",
publisher = "The Optical Society",
number = "8",

}

TY - JOUR

T1 - Enabling fast and high quality LED photoacoustic imaging

T2 - A recurrent neural networks based approach

AU - Anas, Emran Mohammad Abu

AU - Zhang, Haichong K.

AU - Kang, Jin

AU - Boctor, Emad

PY - 2018/8/1

Y1 - 2018/8/1

N2 - Photoacoustic (PA) techniques have shown promise in the imaging of tissue chro-mophores and exogenous contrast agents in various clinical applications. However, the key drawback of current PA technology is its dependence on a complex and hazardous laser system for the excitation of a tissue sample. Although light-emitting diodes (LED) have the potential to replace the laser, the image quality of an LED-based system is severely corrupted due to the low output power of LED elements. The current standard way to improve the quality is to increase the scanning time, which leads to a reduction in the imaging speed and makes the images prone to motion artifacts. To address the challenges of longer scanning time and poor image quality, in this work we present a deep neural networks based approach that exploits the temporal information in PA images using a recurrent neural network. We train our network using 32 phantom experiments; on the test set of 30 phantom experiments, we achieve a gain in the frame rate of 8 times with a mean peak-signal-to-noise-ratio of 35.4 dB compared to the standard technique.

AB - Photoacoustic (PA) techniques have shown promise in the imaging of tissue chro-mophores and exogenous contrast agents in various clinical applications. However, the key drawback of current PA technology is its dependence on a complex and hazardous laser system for the excitation of a tissue sample. Although light-emitting diodes (LED) have the potential to replace the laser, the image quality of an LED-based system is severely corrupted due to the low output power of LED elements. The current standard way to improve the quality is to increase the scanning time, which leads to a reduction in the imaging speed and makes the images prone to motion artifacts. To address the challenges of longer scanning time and poor image quality, in this work we present a deep neural networks based approach that exploits the temporal information in PA images using a recurrent neural network. We train our network using 32 phantom experiments; on the test set of 30 phantom experiments, we achieve a gain in the frame rate of 8 times with a mean peak-signal-to-noise-ratio of 35.4 dB compared to the standard technique.

UR - http://www.scopus.com/inward/record.url?scp=85051289400&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051289400&partnerID=8YFLogxK

U2 - 10.1364/BOE.9.003852

DO - 10.1364/BOE.9.003852

M3 - Article

C2 - 30338160

AN - SCOPUS:85051289400

VL - 9

SP - 3852

EP - 3866

JO - Biomedical Optics Express

JF - Biomedical Optics Express

SN - 2156-7085

IS - 8

M1 - 331886

ER -