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: Jul 5, 2026

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

Kolmogorov-Arnold Guided Local-Global Attention for Medical Image Classification.

Weichao Pan1, Xu Wang2, Chengze Lv2

  • 1School of Computer and Artificial Intelligence, Shandong Jianzhu University, No. 1000 Fengming Road, Ganggou Subdistrict, Jinan, 250101, Shandong Province, People's Republic of China. panweichao01@outlook.com.

Journal of Imaging Informatics in Medicine
|July 1, 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

Double-layer staggered IWFBGs enable multi-dimensional tactile sensing on scalable flexible skin.

Optics express·2026
Same author

TET Loss: A Temperature-Entropy Calibrated Transfer Loss for Reliable Medical Image Classification.

Journal of imaging informatics in medicine·2026
Same author

The Critical Roles of Methyltransferase-Like 3 in Atherosclerosis and Arterial Infarction: Breakthroughs in Mechanism Understanding.

Journal of cellular and molecular medicine·2025
Same author

Pyruvate dehydrogenase complex component B: A Gene Associated with Cuproptosis and Encoding the Beta Subunit of Pyruvate Dehydrogenase Is Involved in the Oxidative Decarboxylation Reaction.

DNA and cell biology·2025
Same author

METTL3-mediated m⁶A methylation in cardiac diseases: pathogenic roles and therapeutic potential.

Cell biology and toxicology·2025
Same author

Formulating Electrophilic Electrolyte for In Situ Stabilization of 4.8 V Li-Rich Batteries with 100% Initial Coulombic Efficiency.

Angewandte Chemie (International ed. in English)·2025

We introduce KA, a novel module for medical image classification that balances local lesion details and global context. This approach enhances representation stability and accuracy with minimal computational cost.

Area of Science:

  • Medical image analysis
  • Deep learning for healthcare
  • Computer vision in medicine

Background:

  • Medical image classification requires integrating fine lesion details and broader anatomical context.
  • Current attention mechanisms often fail to capture both aspects, leading to incomplete feature representations.
  • This limitation impacts the reliability of clinical decisions derived from medical images.

Purpose of the Study:

  • To propose KA, a lightweight, attention-inspired module that balances local and global feature modeling for medical image classification.
  • To enhance the stability and completeness of feature representations in medical imaging tasks.
  • To demonstrate the versatility and effectiveness of KA across different deep learning architectures.

Main Methods:

Keywords:
Attention mechanismKolmogorov–Arnold NetworkMedical image classificationPlug-and-play module

More Related Videos

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

Manual Segmentation of the Human Choroid Plexus Using Brain MRI
04:25

Manual Segmentation of the Human Choroid Plexus Using Brain MRI

Published on: December 15, 2023

Related Experiment Videos

Last Updated: Jul 5, 2026

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

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

Manual Segmentation of the Human Choroid Plexus Using Brain MRI
04:25

Manual Segmentation of the Human Choroid Plexus Using Brain MRI

Published on: December 15, 2023

  • Developed KA, a module based on Kolmogorov-Arnold operators with two complementary pathways: KAN Local Attention Module (KLAM) and KAN Adaptive Mixer (KAM).
  • KLAM utilizes nonlinear modeling within grouped windows to enhance local structures.
  • KAM employs spline-based adaptive fusion to integrate global semantics.
  • Integrated KA into Convolutional Neural Network (CNN), Transformer, and Mamba architectures.

    • Main Results:

    • KA demonstrated consistent performance improvements across three public clinical datasets.
    • The module achieved these gains with limited overhead in parameters and Floating Point Operations (FLOPs).
    • KA effectively balanced local and global feature extraction, leading to more robust representations.

    Conclusions:

    • KA provides a balanced and efficient mechanism for local and global feature extraction in medical image classification.
    • The module's lightweight design and effectiveness make it a valuable addition to various deep learning architectures.
    • KA shows significant potential for improving diagnostic accuracy in clinical settings through enhanced medical image analysis.