MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641056435 A) filed by Dr. D. Jayanthi; Naresh M; Nantha Sridar P; Giridhar J S; and Dr. K. Kiruthika on May 04, 2026, for Secure Multimodal Deep Learning System For Brain And Cardiovascular Health Screening In Iomt.
Inventors include Dr. D. Jayanthi; Naresh M; Nantha Sridar P; Giridhar J S; and Dr. K. Kiruthika Devi.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: A computer-implemented multimodal screening system for indicators of brain and cardiovascular health, integrating heterogeneous consumer-grade physiological data streams - namely a four-channel dry-electrode electroencephalography headband communicating over Bluetooth Low Energy, a still image of a brain magnetic- resonance scan, and a structured portable-document-format report of single-lead electrocardiogram and oscillometric blood-pressure measurements - into a single explainable risk indicator on a mobile device. The system comprises a mobile client that performs on-device digital signal processing on raw electroencephalography samples through a cascaded Butterworth band-pass filter, an adaptive line-noise notch with automatic fifly-or-sixty- hertz mains-frequency selection, a Hanning- windowed fast Fourier transform, and log- normalised band-power computation across the canonical delta, theta, alpha, beta and gamma bands; a window-quality module evaluating each analysis window across exactly six metrics; and a remote inference module hosting a Vision-Transformer-based brain-image classifier and a band-power-matrix mental-state classifier. The mobile client further comprises a portable-document-format parsing module performing private-use-area glyph normalisation prior to regular-expression extraction of cardiovascular fields, a rule-based engine with context-sensitive weights, a brain- cardiovascular correlator producing threshold-gated cross-modality indicators, and a structured eight-section portable-document-format report generator. By transmitting only a compact band- power feature matrix instead of raw electroencephalography samples, the system achieves a reduction in uplink data volume while preserving classification fidelity. The system is designed as an early-indicator screening aid for research and educational purposes and is not represented as a diagnostic medical device.
Disclaimer: Curated by HT Syndication.
