MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641088093 A) filed by Vellore Institute Of Technology on July 19, 2026, for System And Method For Tinyml-Based Multichannel Eeg Seizure Detection And Early Warning.
Inventors include Sivakumar Rajagopal; and Enteamen Yengkhom.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: The present invention relates to a fully edge-based, TinyML-enabled multichannel electroencephalography (EEG) seizure detection and early warning system implemented on a low-cost embedded microcontroller. The system acquires EEG signals from either a 21-channel EEG sensor cap or a graphical user interface (GUI)- based EEG simulator and stores the signals in a sliding window buffer for real-time processing. A lightweight feature extraction module computes computationally efficient EEG features, which are normalized and classified by a TinyML-based Random Forest model executing entirely on an ESP32 microcontroller. Individual channel predictions are aggregated using a multichannel spatial voting mechanism to classify the physiological state as Normal, Pre-seizure, or Seizure while reducing false alarms. The system generates early warnings upon detection of pre-seizure conditions and provides visual and display-based alerts without reliance on cloud computing or external processing infrastructure. The invention offers a portable, low-latency, privacy-preserving, and cost-effective solution for continuous real-time seizure monitoring and early intervention.
Disclaimer: Curated by HT Syndication.
