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 Video

Updated: Nov 5, 2025

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

3.1K

Interpretable and Lightweight 3-D Deep Learning Model for Automated ACL Diagnosis.

YoungSeok Jeon, Kensuke Yoshino, Shigeo Hagiwara

    IEEE Journal of Biomedical and Health Informatics
    |May 18, 2021
    PubMed
    Summary

    We developed a lightweight and interpretable 3D deep neural network for diagnosing anterior cruciate ligament (ACL) tears from knee MRIs. This model achieves high accuracy while being significantly smaller and more explainable than previous methods.

    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

    Investigation of feasibility and diagnostic capability of ultrasound-guided photoacoustic tomography for breast surgical margins.

    Npj imaging·2026
    Same author

    A BIN2-SOG1 molecular module regulates replication stress response independently on ATR.

    Science advances·2026
    Same author

    A longitudinal cross-sectional study of changes in working attitude among primary healthcare workers across the major COVID-19 outbreaks in China.

    Communications medicine·2026
    Same author

    HRRT: hierarchical reinforcement learning for renal replacement therapy decision support.

    NPJ digital medicine·2026
    Same author

    DeepEN: A deep reinforcement learning framework for personalized enteral nutrition in critical care.

    Journal of biomedical informatics·2026
    Same author

    Abdominal fat distribution phenotypes and metabolic risk in premenopausal Asian women: insights beyond general adiposity measures.

    Scientific reports·2026

    Area of Science:

    • Medical Imaging
    • Artificial Intelligence
    • Orthopedics

    Background:

    • Diagnosing anterior cruciate ligament (ACL) tears from knee MRI scans is crucial for orthopedic treatment.
    • Existing deep learning models often prioritize accuracy over practical aspects like interpretability and computational efficiency.
    • Previous methods frequently use large, pre-trained 2D networks and struggle with accurate heatmap generation for model interpretation.

    Purpose of the Study:

    • To propose an interpretable and lightweight 3D deep neural network for diagnosing ACL tears.
    • To address the limitations of existing models regarding explainability and computational cost.
    • To leverage the local and homogeneous characteristics of ACL tear features for improved model design.

    Main Methods:

    More Related Videos

    Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders
    05:49

    Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders

    Published on: November 1, 2024

    1.1K

    Related Experiment Videos

    Last Updated: Nov 5, 2025

    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

    3.1K
    Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders
    05:49

    Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders

    Published on: November 1, 2024

    1.1K
  • Developed a novel 3D deep neural network incorporating attention modules and Gaussian positional encoding to focus on local features.
  • Utilized squeeze modules and reduced convolutional filters to account for feature homogeneity.
  • The model was trained and evaluated on Chiba and Stanford knee MRI datasets.

    • Main Results:

    • The proposed model is highly interpretable, with attention modules accurately highlighting the ACL region without prior location data.
    • The model is exceptionally lightweight, featuring 43K trainable parameters and 7.1 GFLOPs, significantly outperforming prior state-of-the-art in size and computation.
    • Achieved superior diagnostic accuracy, outperforming previous methods with average ROC-AUC scores of 0.983 (Chiba) and 0.980 (Stanford).

    Conclusions:

    • The developed 3D deep neural network offers an effective solution for ACL tear diagnosis, balancing accuracy with interpretability and efficiency.
    • The model's lightweight and interpretable nature makes it suitable for practical clinical applications.
    • This approach demonstrates the potential of tailored network architectures for specific medical imaging tasks.