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: Apr 10, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

878

StableMIL: Entropy-Stabilized Attention-based Multiple Instance Learning for Morphologically Variable Whole Slide

Yinuo Lu, Mingxin Qi, Yao Fu

    IEEE Transactions on Medical Imaging
    |April 8, 2026
    PubMed
    Summary
    This summary is machine-generated.

    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

    A scalable synthesis of N-doped Si nanoparticles for high-performance Li-ion batteries.

    Chemical communications (Cambridge, England)·2016
    Same author

    The effect and action mechanism of resveratrol on the vascular endothelial cell by high glucose treatment.

    Saudi journal of biological sciences·2016
    Same author

    Altered resting state functional connectivity of anterior insula in young smokers.

    Brain imaging and behavior·2016
    Same author

    Curdlan blocks the immune suppression by myeloid-derived suppressor cells and reduces tumor burden.

    Immunologic research·2016
    Same author

    Combined image guided monitoring the pharmacokinetics of rapamycin loaded human serum albumin nanoparticles with a split luciferase reporter.

    Nanoscale·2016
    Same author

    Distinctive effects of CD34- and CD133-specific antibody-coated stents on re-endothelialization and in-stent restenosis at the early phase of vascular injury.

    Regenerative biomaterials·2016
    Same journal

    LLM-enhanced Neuron Segmentation and Reconstruction in Complex Mouse Brain Images.

    IEEE transactions on medical imaging·2026
    Same journal

    Matrixed-Spectrum Decomposition Accelerated Linear Boltzmann Transport Equation Solver for Fast Scatter Correction in Multi-Spectral CT.

    IEEE transactions on medical imaging·2026
    Same journal

    The Ritz Adjoint Method for MRI Pulse Design.

    IEEE transactions on medical imaging·2026
    Same journal

    Physiology-guided Self-supervised Learning for Simultaneous Dual-Tracer PET Separation.

    IEEE transactions on medical imaging·2026
    Same journal

    Informed-Exploration Reinforcement Learning for Automated Virtual Coronary Intervention Planning.

    IEEE transactions on medical imaging·2026
    Same journal

    4D Reconstruction of Fetal Left Ventricle from Echocardiography via 2.5D Radial Segmentation and Graph-Fourier Reconstruction.

    IEEE transactions on medical imaging·2026
    See all related articles

    Computational pathology aggregation methods struggle with Whole Slide Image (WSI) variability. Our StableMIL framework improves attention mechanisms for better generalization in cancer classification and survival prediction.

    Area of Science:

    • Computational pathology
    • Artificial intelligence in medicine
    • Digital pathology

    Background:

    • Aggregating patch features into Whole Slide Image (WSI) representations is vital in computational pathology.
    • Existing methods fail to account for morphological variability in WSIs, leading to attention collapse and misallocation.
    • This limits the generalization and reliability of attention-based models in clinical settings.

    Purpose of the Study:

    • To develop a novel framework, StableMIL, that addresses the limitations of current WSI aggregation strategies.
    • To enhance the robustness of attention mechanisms against variations in patch numbers and spatial distributions.
    • To improve the performance and reliability of computational pathology models for clinical applications.

    Main Methods:

    Related Experiment Videos

    Last Updated: Apr 10, 2026

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
    04:48

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    878
  • Proposed the Entropy-Stabilized Attention-based Multiple Instance Learning (StableMIL) framework.
  • Incorporated an entropy-stabilized attention mechanism for consistent aggregation across varying patch numbers.
  • Utilized Randomly Projected 2D rotary position embedding for robust spatial representation with irregular patch distributions.

    • Main Results:

    • StableMIL effectively handles long instance sequences and out-of-distribution spatial coordinates.
    • Demonstrated consistent outperformance over representative baselines across nine WSI datasets and diverse cancer types.
    • Achieved stable improvements, particularly in survival prediction tasks, across various morphological scenarios.

    Conclusions:

    • StableMIL offers a robust solution for WSI aggregation, overcoming key challenges posed by morphological variability.
    • The framework shows significant potential for real-world clinical applications, enhancing model reliability and performance.
    • StableMIL provides stable improvements in both classification and survival prediction across diverse cancer types.