Dynamic PET Image Denoising Using Deep Convolutional Neural Networks Without Prior Training Datasets

Fumio Hashimoto, Hiroyuki Ohba, Kibo Ote, Atsushi Teramoto, Hideo Tsukada

Research output: Contribution to journalArticle

Abstract

Deep learning has attracted growing interest for application to medical imaging, such as positron emission tomography (PET), due to its excellent performance. Convolutional neural networks (CNNs), a facet of deep learning requires large training-image datasets. This presents a challenge in a clinical setting because it is difficult to prepare large, high-quality patient-related datasets. Recently, the deep image prior (DIP) approach has been devised, based on the fact that CNN structures have the intrinsic ability to solve inverse problems such as denoising without pre-training and do not require the preparation of training datasets. Herein, we proposed the dynamic PET image denoising using a DIP approach, with the PET data itself being used to reduce the statistical image noise. Static PET data were acquired for input to the network, with the dynamic PET images being handled as training labels, while the denoised dynamic PET images were represented by the network output. We applied the proposed DIP method to computer simulations and also to real data acquired from a living monkey brain with 18F-fluoro-2-deoxy-D-glucose (18F-FDG). As a simulation result, our DIP method produced less noisy and more accurate dynamic images than the other algorithms. Moreover, using real data, the DIP method was found to perform better than other types of post-denoising method in terms of contrast-to-noise ratio, and also maintain the contrast-to-noise ratio when resampling the list data to 1/5 and 1/10 of the original size, demonstrating that the DIP method could be applied to low-dose PET imaging. These results indicated that the proposed DIP method provides a promising means of post-denoising for dynamic PET images.

Original languageEnglish
Article number8764327
Pages (from-to)96594-96603
Number of pages10
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 01-01-2019

Fingerprint

Image denoising
Positron emission tomography
Neural networks
Deoxyglucose
Medical imaging
Inverse problems
Glucose
Labels
Brain
Imaging techniques
Computer simulation

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Materials Science(all)
  • Engineering(all)

Cite this

Hashimoto, Fumio ; Ohba, Hiroyuki ; Ote, Kibo ; Teramoto, Atsushi ; Tsukada, Hideo. / Dynamic PET Image Denoising Using Deep Convolutional Neural Networks Without Prior Training Datasets. In: IEEE Access. 2019 ; Vol. 7. pp. 96594-96603.
@article{f5c87b6427524cbea2ea3dc9122387c2,
title = "Dynamic PET Image Denoising Using Deep Convolutional Neural Networks Without Prior Training Datasets",
abstract = "Deep learning has attracted growing interest for application to medical imaging, such as positron emission tomography (PET), due to its excellent performance. Convolutional neural networks (CNNs), a facet of deep learning requires large training-image datasets. This presents a challenge in a clinical setting because it is difficult to prepare large, high-quality patient-related datasets. Recently, the deep image prior (DIP) approach has been devised, based on the fact that CNN structures have the intrinsic ability to solve inverse problems such as denoising without pre-training and do not require the preparation of training datasets. Herein, we proposed the dynamic PET image denoising using a DIP approach, with the PET data itself being used to reduce the statistical image noise. Static PET data were acquired for input to the network, with the dynamic PET images being handled as training labels, while the denoised dynamic PET images were represented by the network output. We applied the proposed DIP method to computer simulations and also to real data acquired from a living monkey brain with 18F-fluoro-2-deoxy-D-glucose (18F-FDG). As a simulation result, our DIP method produced less noisy and more accurate dynamic images than the other algorithms. Moreover, using real data, the DIP method was found to perform better than other types of post-denoising method in terms of contrast-to-noise ratio, and also maintain the contrast-to-noise ratio when resampling the list data to 1/5 and 1/10 of the original size, demonstrating that the DIP method could be applied to low-dose PET imaging. These results indicated that the proposed DIP method provides a promising means of post-denoising for dynamic PET images.",
author = "Fumio Hashimoto and Hiroyuki Ohba and Kibo Ote and Atsushi Teramoto and Hideo Tsukada",
year = "2019",
month = "1",
day = "1",
doi = "10.1109/ACCESS.2019.2929230",
language = "English",
volume = "7",
pages = "96594--96603",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Dynamic PET Image Denoising Using Deep Convolutional Neural Networks Without Prior Training Datasets. / Hashimoto, Fumio; Ohba, Hiroyuki; Ote, Kibo; Teramoto, Atsushi; Tsukada, Hideo.

In: IEEE Access, Vol. 7, 8764327, 01.01.2019, p. 96594-96603.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Dynamic PET Image Denoising Using Deep Convolutional Neural Networks Without Prior Training Datasets

AU - Hashimoto, Fumio

AU - Ohba, Hiroyuki

AU - Ote, Kibo

AU - Teramoto, Atsushi

AU - Tsukada, Hideo

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Deep learning has attracted growing interest for application to medical imaging, such as positron emission tomography (PET), due to its excellent performance. Convolutional neural networks (CNNs), a facet of deep learning requires large training-image datasets. This presents a challenge in a clinical setting because it is difficult to prepare large, high-quality patient-related datasets. Recently, the deep image prior (DIP) approach has been devised, based on the fact that CNN structures have the intrinsic ability to solve inverse problems such as denoising without pre-training and do not require the preparation of training datasets. Herein, we proposed the dynamic PET image denoising using a DIP approach, with the PET data itself being used to reduce the statistical image noise. Static PET data were acquired for input to the network, with the dynamic PET images being handled as training labels, while the denoised dynamic PET images were represented by the network output. We applied the proposed DIP method to computer simulations and also to real data acquired from a living monkey brain with 18F-fluoro-2-deoxy-D-glucose (18F-FDG). As a simulation result, our DIP method produced less noisy and more accurate dynamic images than the other algorithms. Moreover, using real data, the DIP method was found to perform better than other types of post-denoising method in terms of contrast-to-noise ratio, and also maintain the contrast-to-noise ratio when resampling the list data to 1/5 and 1/10 of the original size, demonstrating that the DIP method could be applied to low-dose PET imaging. These results indicated that the proposed DIP method provides a promising means of post-denoising for dynamic PET images.

AB - Deep learning has attracted growing interest for application to medical imaging, such as positron emission tomography (PET), due to its excellent performance. Convolutional neural networks (CNNs), a facet of deep learning requires large training-image datasets. This presents a challenge in a clinical setting because it is difficult to prepare large, high-quality patient-related datasets. Recently, the deep image prior (DIP) approach has been devised, based on the fact that CNN structures have the intrinsic ability to solve inverse problems such as denoising without pre-training and do not require the preparation of training datasets. Herein, we proposed the dynamic PET image denoising using a DIP approach, with the PET data itself being used to reduce the statistical image noise. Static PET data were acquired for input to the network, with the dynamic PET images being handled as training labels, while the denoised dynamic PET images were represented by the network output. We applied the proposed DIP method to computer simulations and also to real data acquired from a living monkey brain with 18F-fluoro-2-deoxy-D-glucose (18F-FDG). As a simulation result, our DIP method produced less noisy and more accurate dynamic images than the other algorithms. Moreover, using real data, the DIP method was found to perform better than other types of post-denoising method in terms of contrast-to-noise ratio, and also maintain the contrast-to-noise ratio when resampling the list data to 1/5 and 1/10 of the original size, demonstrating that the DIP method could be applied to low-dose PET imaging. These results indicated that the proposed DIP method provides a promising means of post-denoising for dynamic PET images.

UR - http://www.scopus.com/inward/record.url?scp=85070320105&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070320105&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2929230

DO - 10.1109/ACCESS.2019.2929230

M3 - Article

AN - SCOPUS:85070320105

VL - 7

SP - 96594

EP - 96603

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8764327

ER -