MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072852 A) filed by Vellore Institute Of Technology on June 12, 2026, for System For Electroencephalogram-Based Seizure Detection Using Attention-Enhanced Neural Network.
Inventors include Dr. Malaya Kumar; T V Manju; Naren Rithvik S; Pranav Atmuri; Battamsetty Vijay Prabhat Sai; and Jai Kishore D.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: ABSTRACT SYSTEM FOR ELECTROENCEPHALOGRAM-BASED SEIZURE DETECTION USING ATTENTION-ENHANCED NEURAL NETWORK A system (100) for electroencephalogram-based seizure detection comprises a processor (102) and a memory (104) storing instructions that cause the processor (102) to receive an electroencephalogram signal (106), transform the electroencephalogram signal (106) into a time-frequency representation (108) using a continuous wavelet transform layer (110), and extract multi-scale features using a parallel pyramidal atrous convolution module (112) comprising multiscale blocks (114) with parallel convolutional branches (116) having different kernel sizes and an aggregation layer (118). The processor (102) refines the extracted features using a convolutional block attention module (120) applying channel attention and spatial attention, models temporal dependencies using a gated recurrent unit layer (122), applies a self- attention layer (124) to assign adaptive importance weights to temporal segments, and classifies the electroencephalogram signal (106) as a seizure state or a non- seizure state based on an output from the self-attention layer (124).
Disclaimer: Curated by HT Syndication.
