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 Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Tool choice matters: Evaluating edgeR vs. DESeq2 for sensitivity, robustness, and cross-study performance.

PloS one·2026
Same author

Integrative Transcriptomic Analysis Reveals Distinct and Shared Host Responses in Dengue and Chikungunya Infections.

International journal of molecular sciences·2026
Same author

Weakly Supervised Teacher-Student Framework with Progressive Pseudo-mask Refinement for Gland Segmentation.

Journal of clinical and translational pathology·2026
Same author

Exploring Ebola virus-associated gene expression through comparative analysis.

Frontiers in genetics·2026
Same author

Predicting protein cascade expression from H&E images.

PLoS computational biology·2026
Same author

Morphology-Aware Prognostic Model for Five-Year Survival Prediction in Colorectal Cancer from H&E Whole-Slide Images: A Study Using Multi-Center Clinical Trial Cohort.

Cancers·2026

Related Experiment Video

Updated: Jul 12, 2025

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

48.1K

Attention2Minority: A salient instance inference-based multiple instance learning for classifying small lesions in

Ziyu Su1, Mostafa Rezapour1, Usama Sajjad1

  • 1Center for Biomedical Informatics, Wake Forest University School of Medicine, Winston-Salem, 27104, USA.

Computers in Biology and Medicine
|October 27, 2023
PubMed
Summary

Salient instance inference MIL (SiiMIL) enhances whole slide image classification by improving the tumor-to-normal ratio in small lesion detection. This weakly-supervised model achieves superior performance and interpretability for pathologists.

Keywords:
Deep learningMultiple instance learningWeakly supervised classificationWhole slide image analysis

More Related Videos

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
13:01

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment

Published on: June 3, 2022

3.8K
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
07:15

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

6.8K

Related Experiment Videos

Last Updated: Jul 12, 2025

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

48.1K
Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
13:01

Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment

Published on: June 3, 2022

3.8K
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
07:15

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

6.8K

Area of Science:

  • Computational pathology
  • Digital pathology
  • Machine learning for medical imaging

Background:

  • Multiple instance learning (MIL) models are successful for whole slide image (WSI) disease classification.
  • Giga-pixel WSI classification is challenged by minute tumor lesions, where small tumor-to-normal area ratios hinder attention mechanisms.

Purpose of the Study:

  • To propose a weakly-supervised MIL model, SiiMIL, for improved WSI classification, particularly for cases with small tumor lesions.
  • To enhance the ability of MIL models to differentiate WSIs with extremely small tumor areas.

Main Methods:

  • Introduced a novel representation learning for histopathology images to identify normal keys.
  • Utilized these keys to select salient instances within WSIs, forming bags with high tumor-to-normal ratios.
  • Employed an attention mechanism for slide-level classification based on the formed bags.

Main Results:

  • SiiMIL improved the tumor-to-normal area ratio in tumor WSIs.
  • Achieved 0.9225 AUC and 0.7551 recall on the Camelyon16 dataset, outperforming existing MIL models.
  • Generated tumor-sensitive attention heatmaps, offering enhanced interpretability for pathologists.

Conclusions:

  • SiiMIL effectively identifies tumor instances, even when comprising less than 1% of a WSI.
  • The model increases the tumor-to-normal instance ratio within bags by two to four times.
  • SiiMIL demonstrates significant potential for accurate and interpretable WSI classification in challenging cases.