MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641099564 A) filed by Sr University on August 18, 2026, for System And Method For Adaptive Explainable Federated Multimodal Lung Cancer Diagnosis.
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 ADAPTIVE EXPLAINABLE FEDERATED MULTIMODAL LUNG CANCER DIAGNOSIS The present disclosure provides an adaptive explainable federated learning system for privacy-preserving multimodal lung-cancer assessment. A plurality of healthcare centers locally process available CT, chest X-ray, and PET/CT images while retaining raw patient images within the respective centers. Each local node preprocesses images, determines available modalities, extracts modality-specific features, and adaptively fuses features without requiring a complete modality set. Local models are trained on institutional data, and privacy-preserving model updates are transmitted to a federated aggregation server for generation and redistribution of an updated global model. During local inference, the system generates an AI-assisted lung-cancer assessment together with an image-based explanation identifying contributory regions and an uncertainty measure indicating prediction reliability. The architecture accommodates heterogeneous imaging resources and missing modalities across participating centers and supports repeated federated model improvement without centralized collection of raw patient medical images, thereby providing privacy- preserving, explainable, and uncertainty-aware clinical decision support.
Disclaimer: Curated by HT Syndication.
