MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202611090959 A) filed by Noida Institute Of Engineering And Technology Niet on July 26, 2026, for Wavelet-Hyperbolic Graph Attention System For Cross-Subject Eeg-Based Mental Stress Detection.
Inventors include Mr. Dhruv Talreja; and Ms. Rashmi Bhardwaj.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The present invention discloses a wavelet-hyperbolic graph attention system (100) and method for cross-subject EEG-based mental stress detection. The system comprises a wearable multi-channel dry-electrode EEG acquisition headset (110), a preprocessing module (120), an adaptive empirical wavelet transform module (130) that derives subject-personalized sub-band boundaries from spectrum extrema, a functional-connectivity graph construction module (140) that builds channel-wise graphs using phase-locking value and weighted phase lag index, a hyperbolic graph convolutional encoder (150) that embeds each graph into a Poincaré ball manifold via Möbius linear layers to preserve hierarchical cortical network structure, a dual-stream cross-modal contrastive temporal transformer (160) that fuses intra-band and inter-band representations, a domain-adversarial subject-invariance module (170), a classification and feedback module (180), and an edge-inference and power management block (190) enabling on-device real-time st
Disclaimer: Curated by HT Syndication.
