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

Accurate and task-agnostic modeling of enzymatic reactions through multimodal relational learning.

Acta pharmaceutica Sinica. B·2026
Same author

Capturing omega sign in the clinical assessment of depression by deep learning.

Translational psychiatry·2026
Same author

VER155008 rescues cognitive impairment in P301S tauopathy mice by promoting HSPA8-mediated lipophagy.

European journal of pharmacology·2026
Same author

Optical metasurfaces for general vision processing on the edge.

Nature·2026
Same author

Author Correction: Large-scale self-supervised video foundation model for intelligent surgery.

NPJ digital medicine·2026
Same author

An AI-Based OCT System to Detect Diabetic Macular Edema: A Prospective Validation and Noninferiority Randomized Clinical Trial.

JAMA·2026

Related Experiment Video

Updated: Oct 26, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.1K

Semi-supervised learning with progressive unlabeled data excavation for label-efficient surgical workflow

Xueying Shi1, Yueming Jin1, Qi Dou2

  • 1Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong.

Medical Image Analysis
|July 29, 2021
PubMed
Summary

This study introduces SurgSSL, a new semi-supervised learning method for surgical workflow recognition. SurgSSL effectively uses unlabeled data to improve model performance, reducing the need for extensive surgeon annotations.

Keywords:
Pseudo label generationSemi-supervised learningSurgical workflow recognitionVideo representation learningVisual temporal dynamic consistency

More Related Videos

Clinical Application of Single-Surgeon, Three-Port, Laparoscopic Resection for Colorectal Cancer with Natural Orifice Specimen Extraction
08:26

Clinical Application of Single-Surgeon, Three-Port, Laparoscopic Resection for Colorectal Cancer with Natural Orifice Specimen Extraction

Published on: March 24, 2023

812
A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment
10:39

A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment

Published on: May 24, 2022

2.5K

Related Experiment Videos

Last Updated: Oct 26, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.1K
Clinical Application of Single-Surgeon, Three-Port, Laparoscopic Resection for Colorectal Cancer with Natural Orifice Specimen Extraction
08:26

Clinical Application of Single-Surgeon, Three-Port, Laparoscopic Resection for Colorectal Cancer with Natural Orifice Specimen Extraction

Published on: March 24, 2023

812
A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment
10:39

A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment

Published on: May 24, 2022

2.5K

Area of Science:

  • Computer-assisted surgery
  • Machine learning in healthcare
  • Surgical workflow analysis

Background:

  • Deep learning models for surgical workflow recognition require large annotated datasets.
  • Manual annotation is time-consuming and requires specialized surgical expertise.
  • Label-efficient methods are crucial for practical application in operating rooms.

Purpose of the Study:

  • To propose a novel two-stage semi-supervised learning method, SurgSSL, for label-efficient surgical workflow recognition.
  • To leverage unlabeled surgical video data to improve model performance and reduce annotation burden.
  • To develop a method that progressively extracts knowledge from unlabeled data.

Main Methods:

  • Proposed a two-stage semi-supervised learning framework named SurgSSL.
  • Introduced an intra-sequence Visual and Temporal Dynamic Consistency (VTDC) scheme for implicit knowledge excavation from unlabeled data.
  • Implemented explicit knowledge excavation using pre-knowledge pseudo-labeling generated by the VTDC-regularized model.

Main Results:

  • SurgSSL significantly outperforms state-of-the-art semi-supervised methods on Cholec80 and M2CAI datasets.
  • Achieved a 10.5% accuracy improvement under limited annotation conditions on the M2CAI dataset.
  • Demonstrated competitive performance with full-data training using only 50% labeled data on the Cholec80 dataset.

Conclusions:

  • SurgSSL effectively reduces the reliance on large annotated datasets for surgical workflow recognition.
  • The proposed method demonstrates superior performance and label efficiency compared to existing approaches.
  • SurgSSL offers a practical solution for advancing computer-assisted surgery applications through improved workflow recognition.