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

Pulmonary Embolism II: Diagnostic Studies and Interprofessional Care01:29

Pulmonary Embolism II: Diagnostic Studies and Interprofessional Care

15
Diagnosing Pulmonary EmbolismDiagnosing pulmonary embolism (PE) involves clinical assessment and advanced imaging tests. The preferred diagnostic tool is the spiral (helical) CT scan or CT angiography (CTA), which uses intravenous contrast media to visualize the pulmonary vasculature and identify emboli.A ventilation-perfusion (V/Q) scan is an alternative for patients unable to receive contrast media. This scan includes both perfusion and ventilation scanning. Perfusion scanning involves...
15

You might also read

Related Articles

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

Sort by
Same author

Non-overtly damaging cold atmospheric plasma induces adaptive epidermal molecular responses associated with integrin alpha 6.

Chemico-biological interactions·2026
Same author

Clinical implementation of 3D deep learning techniques in predicting touch-up lesions for atrial fibrillation patients undergoing cryoablation.

International journal of cardiology. Heart & vasculature·2026
Same author

Results of a Novel Osteotome System for Femoral Stem Extraction in Revision Total Hip Arthroplasty: Technique, Limitations, and Associated Complications.

The Journal of bone and joint surgery. American volume·2026
Same author

Outcomes of Unicompartmental Knee Arthroplasty in Patients Receiving Glucagon-like Peptide 1 Agonist Therapy: A Matched Cohort Study.

Arthroplasty today·2026
Same author

COVID-19 vaccination induces cross-reactive dengue virus antibodies with altered isotype profiles and <i>in vitro</i> antibody-dependent enhancement.

Frontiers in immunology·2026
Same author

Self-Supervised Spatio-Temporal Network for Classifying Lung Tumor in EBUS Videos.

Diagnostics (Basel, Switzerland)·2025

Related Experiment Video

Updated: Jul 29, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

448

Feature-enhanced adversarial semi-supervised semantic segmentation network for pulmonary embolism annotation.

Ting-Wei Cheng1, Yi Wei Chua1, Ching-Chun Huang2

  • 1Department of Mechanical Engineering, College of Engineering, National Yang Ming Chiao Tung University, Hsin-Chu, Taiwan.

Heliyon
|May 22, 2023
PubMed
Summary

This study introduces a semi-supervised learning model for segmenting pulmonary embolism (PE) in CTPA images, improving accuracy on new datasets and reducing labeling costs with unlabeled data.

Keywords:
Computed tomography pulmonary angiogramPulmonary embolismSemantic segmentationSemi-supervised learningUnlabeled images

More Related Videos

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.5K
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.8K

Related Experiment Videos

Last Updated: Jul 29, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

448
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.5K
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.8K

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Computer-Aided Diagnosis

Background:

  • Supervised learning models for pulmonary embolism (PE) detection in computed tomography pulmonary angiogram (CTPA) images require extensive labeled data.
  • Retraining and relabeling are necessary when CTPA images originate from different hospitals, increasing costs and time.
  • A need exists for adaptable segmentation models that can perform accurately across diverse datasets with reduced manual annotation efforts.

Purpose of the Study:

  • To develop and evaluate a feature-enhanced adversarial semi-supervised semantic segmentation model for automated PE lesion annotation in CTPA images.
  • To reduce the reliance on fully labeled datasets by incorporating unlabeled images.
  • To enhance the generalizability and reduce the labor costs associated with PE segmentation models.

Main Methods:

  • A semi-supervised learning approach was implemented, combining a segmentation network and a discriminator network.
  • Feature information from the segmentation network's encoder was integrated into the discriminator to learn label similarities.
  • A modified HRNet-based architecture was employed as the segmentation network to preserve high resolution for detecting small PE lesions.

Main Results:

  • The semi-supervised model achieved mIOU of 0.3510, dice score of 0.4854, and sensitivity of 0.4253 on the NCKUH unlabeled dataset.
  • Fine-tuning on a small set of unlabeled images from CMUH improved performance.
  • Compared to a supervised model, the semi-supervised approach enhanced mIOU from 0.2344 to 0.3721, dice score from 0.3325 to 0.5113, and sensitivity from 0.3151 to 0.4967 on the CMUH dataset.

Conclusions:

  • The proposed semi-supervised model effectively improves the accuracy of PE lesion segmentation on new, unseen datasets.
  • Incorporating a small number of unlabeled images significantly reduces the need for extensive manual labeling.
  • This approach offers a cost-effective and efficient solution for automated PE detection in clinical settings.