Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Multi-modality brain tumor segmentation using dual-attention generative adversarial network.

Reports of practical oncology and radiotherapy : journal of Greatpoland Cancer Center in Poznan and Polish Society of Radiation Oncology·2026
Same author

Development of an End-to-end Phantom and Procedure for Computed Tomography and Cone-beam Computed Tomography-guided Online Adaptive Radiotherapy.

Journal of medical physics·2026
Same author

[Estimation of Air Kerma-area Product Using Real-time Dosimeter in Mobile X-ray Fluoroscopy System].

Nihon Hoshasen Gijutsu Gakkai zasshi·2026
Same author

Radiomic prediction of substantial LVSI in endometrial cancer using reduced field of view DWI- a feasibility study.

European journal of radiology·2026
Same author

Diagnostic performance of susceptibility-weighted imaging for cerebral arteriovenous malformation assessment: An institutional diagnostic accuracy study with structured literature review.

Clinical neurology and neurosurgery·2026
Same author

Facilitating early diagnosis of chronic thromboembolic pulmonary hypertension with dynamic chest radiography: Protocol for a multicenter, assessor-blinded, case-wise randomized superiority reader study (FIND-DCR).

PloS one·2026

Related Experiment Video

Updated: Nov 3, 2025

Author Spotlight: Integrated Photoacoustic, Ultrasound, and Angiographic Tomography (PAUSAT) for NonInvasive Whole-Brain Imaging of Ischemic Stroke
06:45

Author Spotlight: Integrated Photoacoustic, Ultrasound, and Angiographic Tomography (PAUSAT) for NonInvasive Whole-Brain Imaging of Ischemic Stroke

Published on: June 2, 2023

1.9K

Development of attenuation correction methods using deep learning in brain-perfusion single-photon emission computed

Taisuke Murata1, Hajime Yokota2, Ryuhei Yamato3

  • 1Department of Radiology, Chiba University Hospital, Chiba, 260-8677, Japan.

Medical Physics
|June 1, 2021
PubMed
Summary

This study introduces deep learning methods, AutoencoderAC and U-NetAC, for accurate attenuation correction in brain-perfusion SPECT. U-NetAC demonstrated superior performance over ChangAC, generating precise attenuation correction images without extra CT scans.

Keywords:
attenuation correctioncomputed tomographydeep learningradiationsingle-photon emission computed tomography

More Related Videos

Multi-Tracer Studies of Brain Oxygen and Glucose Metabolism Using a Time-of-Flight Positron Emission Tomography-Computed Tomography Scanner
08:36

Multi-Tracer Studies of Brain Oxygen and Glucose Metabolism Using a Time-of-Flight Positron Emission Tomography-Computed Tomography Scanner

Published on: June 7, 2024

469
Studying Metabolic Brain Connectivity Using 2-Deoxy-2-[18F]Fluoro-D-Glucose Dynamic Positron Emission Tomography at the Single-subject Level
07:28

Studying Metabolic Brain Connectivity Using 2-Deoxy-2-[18F]Fluoro-D-Glucose Dynamic Positron Emission Tomography at the Single-subject Level

Published on: January 24, 2025

517

Related Experiment Videos

Last Updated: Nov 3, 2025

Author Spotlight: Integrated Photoacoustic, Ultrasound, and Angiographic Tomography (PAUSAT) for NonInvasive Whole-Brain Imaging of Ischemic Stroke
06:45

Author Spotlight: Integrated Photoacoustic, Ultrasound, and Angiographic Tomography (PAUSAT) for NonInvasive Whole-Brain Imaging of Ischemic Stroke

Published on: June 2, 2023

1.9K
Multi-Tracer Studies of Brain Oxygen and Glucose Metabolism Using a Time-of-Flight Positron Emission Tomography-Computed Tomography Scanner
08:36

Multi-Tracer Studies of Brain Oxygen and Glucose Metabolism Using a Time-of-Flight Positron Emission Tomography-Computed Tomography Scanner

Published on: June 7, 2024

469
Studying Metabolic Brain Connectivity Using 2-Deoxy-2-[18F]Fluoro-D-Glucose Dynamic Positron Emission Tomography at the Single-subject Level
07:28

Studying Metabolic Brain Connectivity Using 2-Deoxy-2-[18F]Fluoro-D-Glucose Dynamic Positron Emission Tomography at the Single-subject Level

Published on: January 24, 2025

517

Area of Science:

  • Medical Imaging
  • Radiology
  • Artificial Intelligence in Medicine

Background:

  • Computed tomography (CT)-based attenuation correction (CTAC) in single-photon emission computed tomography (SPECT) is accurate but requires hybrid SPECT/CT scanners and exposes patients to additional radiation.
  • Previous deep learning methods for generating pseudo-CT images for attenuation correction (AC) have limitations, including misalignment and modality-specific artifacts due to cross-modality transformation.

Purpose of the Study:

  • To develop and evaluate a deep learning-based approach for generating attenuation correction (AC) images in brain-perfusion SPECT directly from non-attenuation-corrected (NAC) images.
  • To compare the performance of the proposed deep learning methods (AutoencoderAC and U-NetAC) against conventional Chang's AC (ChangAC) and CTAC.

Main Methods:

  • Two deep learning models, AutoencoderAC and U-NetAC, were developed using a training dataset of 189 brain-perfusion SPECT patients.
  • A separate test group of 47 patients was used to compare AutoencoderAC, U-NetAC, and ChangAC against CTAC using qualitative visual evaluation and quantitative metrics (NMSE, percentage error).
  • Statistical analyses included the Wilcoxon signed-rank sum test and Bland-Altman analysis.

Main Results:

  • U-NetAC achieved the highest visual evaluation score, indicating superior qualitative accuracy.
  • Quantitative analysis showed U-NetAC had the lowest Normalized Mean Squared Error (NMSE), followed by AutoencoderAC and ChangAC (P < 0.001).
  • ChangAC underestimated counts by 30-40%, while AutoencoderAC and U-NetAC produced mean errors of <1% and maximum errors of 3%, respectively. Bland-Altman analysis revealed biases associated with ChangAC and AutoencoderAC.

Conclusions:

  • Deep learning-based attenuation correction methods (AutoencoderAC and U-NetAC) were successfully developed for brain-perfusion SPECT.
  • Both AutoencoderAC and U-NetAC demonstrated higher accuracy than ChangAC, with U-NetAC showing superior qualitative and quantitative performance.
  • These novel methods generate accurate AC images directly from NAC images, eliminating the need for intermediate pseudo-CT images and additional radiation exposure, though external validation is required.