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: May 13, 2026

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

A Unified Deep Learning Framework for Instance Segmentation Across Diverse Cytological Stains.

Luís Otávio Santos1, Rodrigo de Paula E Silva Ribeiro1, Bibiana Quatrin Tiellet1

  • 1Department of Computer Sciences, Federal University of Santa Catarina, Florianopolis, Brazil.

Cytopathology : Official Journal of the British Society for Clinical Cytology
|May 11, 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

Psychological stress and semen quality impairment: A systematic review.

JBRA assisted reproduction·2026
Same author

The use of an objective method (continuous exosomatic electrodermal activity without external stimuli) to evaluate patients with hyperhidrosis undergoing video-assisted sympathectomy.

Einstein (Sao Paulo, Brazil)·2026
Same author

AI-Driven Fetal Liver Echotexture Analysis: A New Frontier in Predicting Neonatal Insulin Imbalance.

Journal of ultrasound in medicine : official journal of the American Institute of Ultrasound in Medicine·2025
Same author

Prenatal Ultrasonographic Markers of Macrossomia and C-Peptide in Gestational Diabetes Mellitus: A Prospective Cohort Study.

Diagnostics (Basel, Switzerland)·2025
Same author

Continuous evaluation of exosomatic electrodermal activity in patients with primary palmoplantar hyperhidrosis.

Einstein (Sao Paulo, Brazil)·2024
Same author

Case report of the combination of a TRAM flap, lipofilling, and 3-D tattooing after failed implant-based reconstruction: improving aesthetic results.

Journal of surgical case reports·2024

A single deep learning model can accurately segment cells across various cytological stains, improving boundary precision and enabling stain-invariant AI tools for digital pathology. This unified approach simplifies workflows without sacrificing diagnostic accuracy.

Area of Science:

  • Digital Pathology
  • Computational Cytology
  • Artificial Intelligence in Medicine

Background:

  • Cytological analysis relies on diverse staining techniques (e.g., Papanicolaou, Feulgen, AgNOR), necessitating stain-specific image analysis pipelines.
  • Developing a single, versatile model for instance segmentation across these stains is challenging but offers significant workflow advantages.

Purpose of the Study:

  • To assess the feasibility and accuracy of a single instance-segmentation model for robust cell segmentation across multiple cytological stains.
  • To determine if a unified, stain-invariant model can avoid stain-specific pipelines without compromising diagnostic precision.

Main Methods:

  • Consolidated three expert-annotated datasets (Papanicolaou, Feulgen, AgNOR) and standardized formats.
  • Trained instance-segmentation models (Mask R-CNN, Mask2Former, YOLOv11, YOLOv12) under both per-stain and unified multi-stain regimes.
Keywords:
computer‐assistedcytopathologydeep learningimage processinginstance segmentationwhole slide imaging

Related Experiment Videos

Last Updated: May 13, 2026

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

  • Evaluated models using Average Precision at IoU thresholds of 50% (AP50) and 75% (AP75) to assess detection and boundary precision.

    • Main Results:

    • Unified multi-stain training matched or exceeded stain-specific models in performance.
    • The unified approach maintained detection accuracy (AP50) while improving boundary definition (AP75) across diverse stains.
    • Mask2Former demonstrated superior geometric precision (AP75) on complex datasets and the combined dataset.

    Conclusions:

    • A single, unified instance-segmentation model is effective for cytological analysis across different stains.
    • This approach preserves detection accuracy and enhances boundary precision, supporting scalable AI-assisted cytopathology.
    • Transformer-based models like Mask2Former show promise for developing simplified, stain-invariant screening tools in digital pathology.