MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641099566 A) filed by Sr University on August 18, 2026, for System And Method For Privacy-Preserving Federated Deep Learning For Multi-Center Early Lung Cancer Detection.
Inventors include Dr. Ritu Tandon; and Dr. Rupesh Mishra.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: ABSTRACT SYSTEM AND METHOD FOR PRIVACY-PRESERVING FEDERATED DEEP LEARNING FOR MULTI-CENTER EARLY LUNG CANCER DETECTION A privacy-preserving federated deep-learning system for multi-center early lung-cancer detection is disclosed. A plurality of healthcare institutions locally acquire and preprocess computed-tomography and chest X-ray images and train respective deep-learning models without transferring raw patient images to a central repository. Local model information is generated and protected using encryption, differential privacy, secure communication, and secure aggregation. A federated aggregation server combines protected updates from participating institutions and applies a federated optimization process configured to accommodate heterogeneous or Non-IID medical-imaging data. The resulting global lung-cancer model is redistributed to participating institutions for subsequent local training and inference. A local detection module applies the updated model to newly acquired medical images and generates a lung-cancer-related prediction together with a confidence or risk output for clinician decision support. Repeated federated rounds permit collaborative model improvement while retaining sensitive patient imaging data within the originating healthcare institutions.
Disclaimer: Curated by HT Syndication.
