MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202631018763 A) filed by Kaliprasanna Swain, Bhubaneswar, Odisha, on Feb. 19, for 'multimodal federated transformer with multi agent explainability for tauopathy classification.'
Inventor(s) include Sumant Kumar Mohapatra; Kedar Nath Sahu; Kaliprasanna Swain; Ujjwal Sinha; Pralipta Samal; and Rahul Ranjan.
The application for the patent was published on Feb. 27, under issue no. 09/2026.
According to the abstract released by the Intellectual Property India: "The present invention (100) relates to the field of medical image analysis and artificial intelligence for neurodegenerative disease diagnosis. It addresses limitations of existing tauopathy diagnostic systems that rely on single-modality inputs and lack disease-specific modeling, graph-based biomarker reasoning, federated learning support, and comprehensive explainability. The invention provides a unified multimodal transformer architecture integrating a Residual Deformation-Aware 3D Transformer (RDAT) for MRI, a Graph-RoBERTa encoder for ROI biomarkers (101), and a TabO2-MLP encoder for metadata (101). A Cross-Tau Fusion HyperAttention module (102) dynamically weights tauopathy-specific features, while a Multi-Agent Neurodiagnostic Layer (103) combines morphology, biomarker, and clinical reasoning agents to form consensus predictions. The system incorporates federated learning for privacy-preserving training and an explainability suite using Grad-CAM3D, SHAP, and LIME. Principal use is automated, explainable classification of tauopathies (PSP, MSA, CBD, CTRL) in clinical and research settings."
Disclaimer: Curated by HT Syndication.
