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

Frequency-domain multi-scale hybrid attention for pathological image classification.

Yang Zhang1, Junjie Li1, Qiushi Wang2

  • 1College of Computer and Information Sciences, Chongqing Normal University, Chongqing, 401331, China.

Scientific Reports
|June 23, 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

An early prediction model of steroid dependence or frequent relapses in nephrotic children: a multicentre retrospective cohort study in China.

Clinical kidney journal·2026
Same author

African Swine Fever Virus R238L and R298L Disrupt Lung Cell Collagen Formation and Cell Adhesion Pathway by Targeting Transcription Factors Containing zf-C2H2 Domain.

Veterinary sciences·2026
Same author

Identification of Regulators for Antigen-Specific CD8<sup>+</sup> T Cells in African Swine Fever Virus-Restored Pigs.

Veterinary sciences·2025
Same author

Procyanidin-B1-Enriched <i>Cyperus esculentus</i> Extract Regulates Anti-Inflammatory Pathways in Chicken Macrophages Cells Through Transcription Factor STAT2 and HIF1A.

Animals : an open access journal from MDPI·2025
Same author

Machine learning model for predicting severe infection in children with idiopathic nephrotic syndrome: multicenter retrospective study.

Italian journal of pediatrics·2025
Same author

Recombinant rSF toxoid enhances efficacy of bacterin-inactivated vaccine against F18+ Escherichia coli-induced edema disease in swine.

BMC veterinary research·2025
This summary is machine-generated.

This study introduces a novel Multi-scale Frequency-domain Hybrid Attention mechanism (MFHA) for enhanced pathological image classification. MFHA improves cancer diagnosis by effectively fusing global context and local details for greater accuracy.

Area of Science:

  • Digital pathology
  • Medical image analysis
  • Computational oncology

Background:

  • Pathological image classification is vital for cancer diagnosis and subtyping.
  • Significant heterogeneity and complex textures in pathological images pose challenges.
  • Existing methods struggle to integrate global context with local details effectively.

Purpose of the Study:

  • To propose a novel Multi-scale Frequency-domain Hybrid Attention mechanism (MFHA) for improved pathological image classification.
  • To address limitations in exploiting local details and fusing frequency-domain information.
  • To enhance the joint modeling of global structure and local texture.

Main Methods:

  • Utilized wavelet transform for image decomposition into low-frequency (global) and high-frequency (local) subbands.
Keywords:
AttentionComputational pathologyFeature enhancementMulti-scaleWavelet transform

Related Experiment Videos

  • Integrated multi-scale convolutions with subband fusion to enhance high-frequency feature representation.
  • Introduced a spatial attention module using cosine similarity and multi-dimensional statistics for robustness.

    • Main Results:

    • The proposed MFHA method demonstrated superior performance over baseline methods on multiple pathological datasets.
    • Achieved accuracy gains of 1.97% and 2.71% in pathological image classification.
    • Significantly improved feature robustness and discriminability.

    Conclusions:

    • MFHA offers a novel approach for co-modeling pathological image features.
    • The method effectively boosts classification accuracy and robustness in digital pathology.
    • This work provides a valuable tool for early cancer diagnosis and precise subtyping.