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: Jan 9, 2026

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.3K

Data-Efficient Deep Learning Framework for Urolithiasis Detection Using Transfer and Self-Supervised Learning.

Jae-Seoung Kim1, Sung-Jong Eun2

  • 1Core Research & Development Center, Korea University Ansan Hospital, Ansan, Korea.

International Neurourology Journal
|December 8, 2025
PubMed
Summary
This summary is machine-generated.

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

Evolutionary Frontiers in Urology: Expanding Clinical Care Through Devices and Intelligence.

International neurourology journal·2026
Same author

From Machine Learning to Generative Artificial Intelligence in Urology: Technological Evolution and Future Perspectives.

International neurourology journal·2026
Same author

Impact of Simulated Radiation Dose Reduction on Deep Learning-Based Renal Segmentation Performance: A Simulation Study Using the KiTS21 (2021 Kidney and Kidney Tumor Segmentation Challenge) Dataset.

International neurourology journal·2026
Same author

Changes and Evolution of Artificial Specialized Intelligence in Urology.

International neurourology journal·2025
Same author

Radiomics Reproducibility in Prostate Cancer Diagnosis Based on PROSTATEx.

International neurourology journal·2025
Same author

A lightweight YOLOv8-based model for gastric cancer detection.

Computers in biology and medicine·2025

This study introduces a data-efficient framework combining self-supervised learning (SSL) and transfer learning (TL) for accurate urolithiasis detection using limited computed tomography (CT) scans. The approach significantly improves diagnostic performance in small-data clinical settings.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Urology

Background:

  • Deep learning for urolithiasis detection typically requires large labeled datasets, which are often unavailable in clinical practice.
  • Limited and partially labeled computed tomography (CT) scans restrict the generalizability of conventional supervised models.
  • Data scarcity poses a significant challenge for developing robust AI diagnostic tools in urology.

Purpose of the Study:

  • To propose a data-efficient framework for accurate urolithiasis detection from small CT datasets.
  • To integrate self-supervised learning (SSL) and transfer learning (TL) to overcome data limitations.
  • To develop a generalizable and resource-efficient diagnostic model for urological imaging.

Main Methods:

  • A SimCLR-based SSL framework with a ResNet50 backbone was used to learn feature representations from 100 unlabeled abdominal CT scans.
Keywords:
Computed tomographyDeep learningTransfer learningUrolithiasis

Related Experiment Videos

Last Updated: Jan 9, 2026

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

2.3K
  • The pretrained encoder was fine-tuned using labeled data via transfer learning (TL) with a linear classifier.
  • 5-fold cross-validation was employed to evaluate model performance using accuracy, precision, recall, F1-score, and AUC.

    • Main Results:

    • The combined SSL+TL model achieved superior performance with an AUC of 0.95 and an F1-score of 0.91.
    • The proposed model significantly outperformed models trained with random initialization and TL-only approaches.
    • SSL pretraining effectively learned robust and transferable representations from limited data.

    Conclusions:

    • The developed framework demonstrates the feasibility of AI-based urolithiasis detection in resource-constrained clinical environments.
    • Combining SSL and TL effectively addresses data scarcity issues in medical imaging.
    • This approach provides a foundation for creating more generalizable and efficient AI diagnostic models for urological conditions.