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 Experiment Videos

TriNet-MoE: One New Neural Network Framework Based on Mixture of Experts for CXR-Based COVID-19 Detection.

Chunhua Zhu1,2,3, Shuzhi Yang4,5,6, Xue Li1,2,7

  • 1Key Laboratory of Grain Information Processing and Control, Henan University of Technology, Ministry of Education, Zhengzhou, 450001, China.

Journal of Imaging Informatics in Medicine
|June 11, 2026
PubMed
Summary

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

Evidence-Based Guidelines for the Diagnosis and Treatment of Pediatric CKD-Mineral and Bone Disorder (Version 2024).

Kidney international reports·2026
Same author

Executive Summary of Evidence-Based Guidelines for the Diagnosis and Treatment of Pediatric CKD-Mineral and Bone Disorder (Version 2024).

Kidney international reports·2026
Same author

Detection of duck adenovirus 3 using RAA-CRISPR/Cas12a based lateral flow dipstick method.

Frontiers in microbiology·2026
Same author

Development of RT-RAA-CRISPR/Cas12a-Based Rapid Visual Detection Assay for Pigeon Rotavirus A.

Microorganisms·2026
Same author

Circular RNA profiling uncovers regulators of early body color transition in juvenile leopard coral grouper (Plectropomus leopardus).

BMC genomics·2026
Same author

Impact of Slaughter-Induced Stress on Biochemical Deterioration and Flavor Stability of Greater Amberjack (Seriola dumerili) Fillets During Iced Storage.

Journal of food science·2026

A new TriNet-MoE deep learning model improves COVID-19 detection from chest X-rays by integrating local and global features. This framework achieves high accuracy, outperforming existing methods for lung disease screening.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Computer Vision

Background:

  • Chest X-ray (CXR) is crucial for lung disease screening, including COVID-19.
  • Current deep learning models struggle to combine local lesion details with global context for accurate CXR analysis.
  • Limited discrimination in existing methods hinders effective COVID-19 detection and pneumonia recognition.

Purpose of the Study:

  • To develop an advanced deep learning framework for enhanced CXR-based COVID-19 detection.
  • To address the challenge of integrating fine-grained local features with global spatial context in medical imaging.
  • To improve the diagnostic accuracy of lung disease detection using artificial intelligence.

Main Methods:

  • Proposed a novel triple neural network framework named TriNet-MoE (Mixture of Experts).
Keywords:
COVID-19 detectionChest X-rayImage recognitionMixture of experts

Related Experiment Videos

  • Integrated ResNet34, ResNet50 for local features, and Vision Transformer for global context.
  • Introduced Cross-Attention Synergy (CA-Synergy) for hierarchical feature interaction and MoE-Intelligent Linked Feature Collaborator (MoE-ILFC) for adaptive fusion.

    • Main Results:

    • Achieved high accuracies of 98.73% on DLAI3 and 98.78% on COVIDx datasets.
    • Consistently outperformed representative baseline models under identical experimental conditions.
    • Demonstrated stable performance improvements and strong generalization capabilities across datasets and domain shifts.

    Conclusions:

    • TriNet-MoE effectively integrates local and global features for robust COVID-19 detection from CXRs.
    • The proposed framework offers superior diagnostic performance compared to existing methods.
    • Visualization analyses confirm the model's ability to leverage informative features for accurate lung disease identification.