What question did this study set out to answer?

To develop an adaptive federated learning framework fostering high-accuracy diagnostics in medical imaging across institutions.

May 7, 2026Scientific ReportsOpen Access

Adaptive homomorphic federated learning framework for multi-institutional medical imaging with optimized diagnostic accuracy

View Full Paper

Ask AI

Bookmark

Authors

Framework improves diagnostic accuracy for lung nodules in multi-institutional settings, suggesting better AI integration in healthcare.

To develop an adaptive federated learning framework fostering high-accuracy diagnostics in medical imaging across institutions.
Utilized NASFL framework combining multi-level homomorphic encryption and stochastic differential privacy.
Incorporated a transformer-guided ResNet for feature fusion from CT and X-ray data.
Employed top-k gradient compression and adaptive learning rates to enhance communication efficiency.
Tested on multi-institutional datasets including NIH Chest X-ray14 and LIDC-IDRI.
Achieved 99.6% diagnostic accuracy across multiple datasets.
Reduced convergence time to approximately 65 communication rounds.
Demonstrated robust performance in heterogeneous data environments.
Set a new benchmark for scalable medical diagnostics with strong privacy assurance.

Usha et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

May 7, 2026Scientific ReportsOpen Access

View Full Paper

Ask AI

Bookmark

Authors

Framework improves diagnostic accuracy for lung nodules in multi-institutional settings, suggesting better AI integration in healthcare.

To develop an adaptive federated learning framework fostering high-accuracy diagnostics in medical imaging across institutions.
Utilized NASFL framework combining multi-level homomorphic encryption and stochastic differential privacy.
Incorporated a transformer-guided ResNet for feature fusion from CT and X-ray data.
Employed top-k gradient compression and adaptive learning rates to enhance communication efficiency.
Tested on multi-institutional datasets including NIH Chest X-ray14 and LIDC-IDRI.
Achieved 99.6% diagnostic accuracy across multiple datasets.
Reduced convergence time to approximately 65 communication rounds.
Demonstrated robust performance in heterogeneous data environments.
Set a new benchmark for scalable medical diagnostics with strong privacy assurance.

Usha et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark