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

REMIL-IBD: Region-filtered multiple instance learning for interpretable slide-level grading of inflammatory bowel disease.

Journal of pathology informatics·2026
Same author

Assessing perceived needs for telepathology implementation in Colombia: a baseline study from Red GLORIA.

Frontiers in digital health·2026
Same author

Proceedings of the first international NMIBC consensus classification meeting.

Bladder cancer (Amsterdam, Netherlands)·2026
Same author

Tactile stimulation and newborn heart rate responses in the delivery room: an observational video study.

Resuscitation·2026
Same author

CARMN fusions and PDGFRB mutations highlight key differences among cutaneous tumors with myopericytic differentiation.

The Journal of investigative dermatology·2026
Same author

The impact of tissue detection on diagnostic artificial intelligence algorithms in prostate digital pathology.

Scientific reports·2026

Related Experiment Video

Updated: Jun 11, 2025

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

3.6K

Equipping computational pathology systems with artifact processing pipelines: a showcase for computation and

Neel Kanwal1, Farbod Khoraminia2, Umay Kiraz3,4

  • 1Department of Electrical Engineering and Computer Science, University of Stavanger, 4021, Stavanger, Norway. neel.kanwal@uis.no.

BMC Medical Informatics and Decision Making
|October 7, 2024
PubMed
Summary

This study introduces a Mixture of Experts (MoE) approach for detecting artifacts in whole slide images (WSIs), improving computational pathology (CPATH) reliability. The DCNNs-based MoE achieved high accuracy, offering a robust solution for automated cancer diagnosis.

Keywords:
Computational pathologyDeep learningHistological artifactsMixture of expertsVision transformerWhole slide images

More Related Videos

Establishment and Analysis of Tumor Slice Explants As a Prerequisite for Diagnostic Testing
11:00

Establishment and Analysis of Tumor Slice Explants As a Prerequisite for Diagnostic Testing

Published on: November 29, 2018

14.2K
Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis
07:45

Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis

Published on: October 25, 2024

331

Related Experiment Videos

Last Updated: Jun 11, 2025

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

3.6K
Establishment and Analysis of Tumor Slice Explants As a Prerequisite for Diagnostic Testing
11:00

Establishment and Analysis of Tumor Slice Explants As a Prerequisite for Diagnostic Testing

Published on: November 29, 2018

14.2K
Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis
07:45

Author Spotlight: Standardizing Mouse In Vivo PET Imaging with Body Conforming Molds and Automated Analysis

Published on: October 25, 2024

331

Area of Science:

  • Computational pathology and digital histopathology.
  • Application of deep learning (DL) in medical image analysis.

Background:

  • Histopathology is crucial for cancer diagnosis, relying on microscopic examination of tissue slides.
  • Tissue processing introduces artifacts into whole slide images (WSIs), potentially leading to misdiagnoses by deep learning (DL) algorithms.
  • Automated artifact detection is essential for reliable computational pathology (CPATH) systems.

Purpose of the Study:

  • To develop and evaluate a Mixture of Experts (MoE) scheme for detecting five common artifacts in WSIs.
  • To assess the performance and computational trade-offs of different DL pipelines for artifact detection.
  • To ensure reliable CPATH predictions and provide quality control for digital pathology.

Main Methods:

  • Proposed a Mixture of Experts (MoE) scheme using independent binary DL models as experts for artifact detection.
  • Ensembled expert predictions using a fusion mechanism and applied probabilistic thresholding to enhance sensitivity.
  • Developed and evaluated four DL pipelines (two MoEs, two multiclass models with DCNNs and ViTs) on diverse datasets, including out-of-distribution data.

Main Results:

  • DCNNs-based MoE and ViTs-based MoE schemes outperformed simpler multiclass models.
  • The best-performing pipeline, MoE using MobileNet DCNNs, achieved 86.15% F1 and 97.93% sensitivity on unseen data.
  • Expert evaluation confirmed the diagnostic usability of the DCNN-based MoE scheme, with a Cohen Kappa of 0.82 for artifact detection and preservation of artifact-free regions.

Conclusions:

  • The proposed artifact detection pipeline enhances CPATH reliability and offers quality control.
  • MoE with DCNNs demonstrated superior performance for artifact detection, balancing accuracy and computational cost.
  • The study highlights the trade-off between performance and complexity, emphasizing that no single DL solution fits all applications.