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

Federated Learning with Differential Privacy for Ultrasound Breast Cancer Classification: An Empirical Study.

Nursultan Makhanov1, Beibit Abdikenov1, Tomiris Zhaksylyk1

  • 1Science and Innovation Center "Artificial Intelligence", Astana IT University, Astana 010000, Kazakhstan.

Journal of Imaging
|May 26, 2026
PubMed
Summary

Related Concept Videos

Imaging Studies II: Ultrasonography01:24

Imaging Studies II: Ultrasonography

IntroductionUltrasonography, or renal ultrasound, is a noninvasive medical imaging technique that uses high-frequency sound waves to visualize the kidneys, ureters, bladder, and surrounding tissues.Indications for Urinary System UltrasonographyUrinary system ultrasonography is indicated in various clinical scenarios, such as:Kidney Stones (Urolithiasis): To detect and monitor the size and presence of kidney or urinary tract stones.Hydronephrosis: To assess the dilation of the renal pelvis and...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Assessing CNNs and LoRA-Fine-Tuned Vision-Language Models for Breast Cancer Histopathology Image Classification.

Journal of imaging·2026
Same author

Federated Self-Supervised Few-Shot Face Recognition.

Journal of imaging·2025
Same author

Enhancing Breast Lesion Detection in Mammograms via Transfer Learning.

Journal of imaging·2025
Same author

Innovative Multi-View Strategies for AI-Assisted Breast Cancer Detection in Mammography.

Journal of imaging·2025

Federated learning (FL) enables privacy-preserving breast cancer classification using deep learning on ultrasound images. Models achieve high accuracy comparable to centralized methods, even with differential privacy, though diagnostic tasks show performance degradation.

Area of Science:

  • Artificial Intelligence
  • Medical Imaging
  • Oncology

Background:

  • Deep learning holds promise for breast cancer classification in medical images.
  • Privacy regulations (HIPAA, GDPR) hinder centralized data aggregation for AI model training.
  • Federated learning (FL) offers a decentralized approach to address data privacy concerns.

Purpose of the Study:

  • To empirically evaluate federated learning (FL) for breast cancer classification in ultrasound images.
  • To compare the performance of various deep learning architectures and FL algorithms under privacy constraints.
  • To provide guidance on architecture selection, FL algorithm choice, and privacy-utility trade-offs.

Main Methods:

  • Systematic comparison of seven deep learning architectures (ResNet-50, VGG16, VGG19, DenseNet-121, MobileNetV2, Vision Transformer, CoAtNet).
Keywords:
breast cancerdifferential privacyfederated learningimage classificationmedical imaging

Related Experiment Videos

  • Evaluation across three FL algorithms (FedAvg, FedProx, FedOpt) with client-side differential privacy (DP).
  • Simulated federation of eight institutions across three classification scenarios.

    • Main Results:

    • Federated models achieved performance comparable to centralized baselines (e.g., 98.52% accuracy for normal/abnormal screening).
    • Vision Transformer (ViT-small) and DenseNet-121 outperformed centralized models in some configurations.
    • Strong differential privacy (η=2.0) maintained screening accuracy (>82%) but degraded diagnostic task performance (best 68.42%).

    Conclusions:

    • Federated learning is a viable approach for privacy-preserving breast cancer classification.
    • Architecture and algorithm choices impact performance and privacy-utility trade-offs.
    • Further research is needed to address challenges for clinical deployment of privacy-preserving AI in breast cancer diagnosis.