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: Jan 9, 2026

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

3.3K

Expert-Like Reparameterization of Heterogeneous Pyramid Receptive Fields in Efficient CNNs for Fair Medical Image

Xiao Wu, Xiaoqing Zhang, Zunjie Xiao

    IEEE Transactions on Medical Imaging
    |December 8, 2025
    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

    Improving the Electrocatalytic Activity and Durability of the La<sub>0.6</sub>Sr<sub>0.4</sub>Co<sub>0.2</sub>Fe<sub>0.8</sub>O<sub>3-δ</sub> Cathode by Surface Modification.

    ACS applied materials & interfaces·2018
    Same author

    Oriented electron transmission in polyoxometalate-metalloporphyrin organic framework for highly selective electroreduction of CO<sub>2</sub>.

    Nature communications·2018
    Same author

    Immune Checkpoint Inhibition Overcomes ADCP-Induced Immunosuppression by Macrophages.

    Cell·2018
    Same author

    Impact of Strain-Induced Changes in Defect Chemistry on Catalytic Activity of Nd<sub>2</sub>NiO<sub>4+δ</sub> Electrodes.

    ACS applied materials & interfaces·2018
    Same author

    Curcumin induces apoptosis and inhibits angiogenesis in murine malignant mesothelioma.

    International journal of oncology·2018
    Same author

    Deletion of SMARCA4 impairs alveolar epithelial type II cells proliferation and aggravates pulmonary fibrosis in mice.

    Genes & diseases·2018

    This study introduces Expert-Like Reparameterization of Heterogeneous Pyramid Receptive Fields (ERoHPRF) to improve medical image classification by capturing diverse lesion characteristics and enhancing prediction fairness, outperforming existing methods.

    Area of Science:

    • Medical Image Analysis
    • Computer Vision
    • Artificial Intelligence

    Background:

    • Convolutional Neural Networks (CNNs) often use single or limited receptive fields (RFs), struggling with diverse medical lesion characteristics.
    • Existing CNNs face challenges in efficient feature representation for imbalanced medical image classification and exhibit biases in predictions.

    Purpose of the Study:

    • To develop a novel concept, Expert-Like Reparameterization of Heterogeneous Pyramid Receptive Fields (ERoHPRF), to enhance medical image classification performance and fairness.
    • To address limitations in capturing diverse lesion characteristics and mitigate prediction bias in CNNs for medical applications.

    Main Methods:

    • ERoHPRF utilizes a heterogeneous pyramid RF bag with multiple kernel sizes to effectively capture lesion characteristics of varying significance.

    Related Experiment Videos

    Last Updated: Jan 9, 2026

    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

    3.3K
  • An expert-like structural reparameterization technique merges parameters in a two-stage strategy for computational efficiency and inference speed.
  • ERoHPRF was integrated into mainstream efficient CNN architectures for evaluation.

    • Main Results:

    • ERoHPRF demonstrated superior performance in medical image classification, effectively capturing diverse lesion features.
    • The method significantly improved the fairness of CNN predictions, reducing bias in medical diagnostic tasks.
    • Experiments confirmed ERoHPRF offers a better trade-off between classification accuracy, fairness, and computational overhead compared to state-of-the-art methods.

    Conclusions:

    • ERoHPRF offers an effective approach to simultaneously boost performance and fairness in medical image classification.
    • The proposed method mimics multi-expert consultation to better handle complex and imbalanced medical image datasets.
    • ERoHPRF presents a promising advancement for reliable and unbiased AI-driven medical diagnosis.