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

Synthetic data generation: challenges and perspectives for gastrointestinal medicine.

Nature reviews. Gastroenterology & hepatology·2026
Same author

When silence is safer: a review and decision-theoretic framework for LLM abstention in healthcare.

NPJ digital medicine·2026
Same author

Diverse image generation with diffusion models and cross class label learning for polyp classification.

Scientific reports·2026
Same author

Colonoscopy in elderly patients: diagnostic yield, pre-referral assessment and quality indicators in two Swedish cohorts.

Scandinavian journal of gastroenterology·2026
Same author

Positional differences in peak- and accumulated-training load relative to match load in highly-trained female football players.

Frontiers in sports and active living·2026
Same author

Endoscopist as a key determinant of patient pain: a sigmoidoscopy screening model for unbiased assessment.

Scandinavian journal of gastroenterology·2026

Related Experiment Video

Updated: Nov 23, 2025

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
07:13

Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

Published on: April 18, 2025

355

A Comprehensive Study on Colorectal Polyp Segmentation With ResUNet++, Conditional Random Field and Test-Time

Debesh Jha, Pia H Smedsrud, Dag Johansen

    IEEE Journal of Biomedical and Health Informatics
    |January 5, 2021
    PubMed
    Summary

    Improving colorectal cancer detection, this study enhances the ResUNet++ model with Conditional Random Field (CRF) and Test-Time Augmentation (TTA). These advancements significantly boost polyp detection accuracy, especially for small, flat lesions, reducing missed abnormalities during colonoscopy.

    More Related Videos

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

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    654
    Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
    06:46

    Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery

    Published on: September 27, 2024

    506

    Related Experiment Videos

    Last Updated: Nov 23, 2025

    Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
    07:13

    Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model

    Published on: April 18, 2025

    355
    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
    04:48

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    654
    Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
    06:46

    Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery

    Published on: September 27, 2024

    506

    Area of Science:

    • Medical Imaging
    • Artificial Intelligence
    • Gastroenterology

    Background:

    • Colonoscopy is the gold standard for colorectal cancer detection but suffers from high miss-rates.
    • Computer-Aided Diagnosis (CADx) systems using machine learning can improve lesion detection during endoscopy.
    • The ResUNet++ architecture previously showed improved performance over U-Net and ResUNet for polyp segmentation.

    Purpose of the Study:

    • To enhance the prediction performance of the ResUNet++ architecture for polyp segmentation.
    • To evaluate the impact of Conditional Random Field (CRF) and Test-Time Augmentation (TTA) on ResUNet++.
    • To assess the model's generalization capability across diverse polyp datasets and its effectiveness on challenging small, flat polyps.

    Main Methods:

    • Implemented Conditional Random Field (CRF) and Test-Time Augmentation (TTA) with the ResUNet++ architecture.
    • Conducted extensive evaluations on six public datasets (Kvasir-SEG, CVC-ClinicDB, CVC-ColonDB, ETIS-Larib Polyp DB, ASU-Mayo Clinic Colonoscopy Video Database, CVC-VideoClinicDB).
    • Performed cross-dataset evaluations to test generalization and analyzed performance on a curated subset of 196 small, flat polyps.

    Main Results:

    • The integration of CRF and TTA significantly improved ResUNet++ performance on polyp segmentation tasks.
    • Enhanced accuracy was observed in both same-dataset and cross-dataset evaluations.
    • The refined model demonstrated strong performance in detecting small (less than 10mm), flat, or sessile polyps, which are frequently missed.

    Conclusions:

    • Conditional Random Field (CRF) and Test-Time Augmentation (TTA) effectively enhance the ResUNet++ architecture for polyp segmentation.
    • The improved model shows promise for real-world clinical application by increasing the detection rate of difficult-to-identify polyps.
    • Further investigation into these methods is warranted for clinical practice to reduce colorectal cancer miss-rates.