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 Videos

Stream-based active learning for surgical AI.

Gregor Just1,2, Alexander C Jenke3, Antonia Kraneis4

  • 1Department of Translational Surgical Oncology, National Center for Tumor Diseases (NCT), NCT/UCC Dresden, a partnership between DKFZ, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, and Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Dresden, Germany. gregor.just@nct-dresden.de.

International Journal of Computer Assisted Radiology and Surgery
|June 15, 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

Appendix300: Surgical video and patient metadata of 330 laparoscopic appendectomy cases from five institutions.

Scientific data·2026
Same author

Decentralized, privacy-preserving surgical video analysis with Swarm Learning.

medRxiv : the preprint server for health sciences·2025
Same author

Appendix300: A multi-institutional laparoscopic appendectomy video dataset for computational modeling tasks.

medRxiv : the preprint server for health sciences·2025
Same author

AutoFRS: an externally validated, annotation-free approach to computational preoperative complication risk stratification in pancreatic surgery - an experimental study.

International journal of surgery (London, England)·2025
Same author

6G networks for the operating room of the future.

Progress in biomedical engineering (Bristol, England)·2024
Same author

A surgical activity model of laparoscopic cholecystectomy for co-operation with collaborative robots.

Surgical endoscopy·2024
Same journal

Correspondence-free local-to-global liver deformation correction via implicit neural representation and biomechanical model.

International journal of computer assisted radiology and surgery·2026
Same journal

BronchoLumen: analysis of recent YOLO-based architectures for real-time bronchial orifice detection in video bronchoscopy.

International journal of computer assisted radiology and surgery·2026
Same journal

Model-guided medicine for early diagnosis of transthyretin-associated cardiac amyloidosis using multimodal data integration and standardized interoperable models (the CRONOS-ATTR study).

International journal of computer assisted radiology and surgery·2026
Same journal

Electromagnetic navigation for femoral osteotomy using high-accuracy X-ray-to-CT registration.

International journal of computer assisted radiology and surgery·2026
Same journal

Modular instrument actuation unit for robotic-assisted systems in laparoscopic surgery.

International journal of computer assisted radiology and surgery·2026
Same journal

Pose-aware deep perceptual similarity for iterative 2D/3D registration of knee joints using contrastive learning.

International journal of computer assisted radiology and surgery·2026
See all related articles
This summary is machine-generated.

Continuous deployment of surgical AI requires efficient annotation. Stream-based Active Learning enables models to adapt during surgery, improving performance with limited expert time and enhancing generalization.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Surgical Technology

Background:

  • Continuous deployment of AI in surgery requires adaptive models that evolve with clinical practice.
  • Efficient data annotation is crucial for sustained model improvement, balancing expert time with the need for informative data.
  • Intraoperative annotation offers opportunities for real-time feedback and model refinement.

Purpose of the Study:

  • To propose and evaluate a Stream-based Active Learning pipeline for continuous deployment in surgical AI.
  • To enable online selection of informative training data during surgical procedures.
  • To adapt AI models and mitigate distribution shifts through continuous learning.

Main Methods:

  • Utilized VeSSAL, a Stream-based Active Learning method leveraging model gradients, uncertainty, and data diversity.
Keywords:
Active learningIntraoperative annotationModel lifecycleSurgical AI

Related Experiment Videos

  • Incorporated an Adversarial Autoencoder (AAE) loss component for latent feature normalization to improve sample selection.
  • Enabled online selection of informative training data during live surgical procedures.

    • Main Results:

    • The enhanced VeSSAL achieved a stable labeling ratio with a 12% standard deviation.
    • Demonstrated an improvement in macro F1 score by up to 7.7% after 15 surgeries compared to baseline strategies.
    • Evaluated performance on a private dataset and the Cholec80 benchmark dataset.

    Conclusions:

    • Intraoperative Stream-based Active Learning is feasible and label-efficient for continuous deployment in surgical AI.
    • Gradient-based Active Learning combined with AAE for feature normalization enhances model generalization and sample selection.
    • This approach provides a practical pathway for continuously learning surgical AI systems.