MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641105749 A) filed by S S Imambi on September 02, 2026, for A Multi-Domain Transformer-Based Framework For Eeg Artifact Detection And Reconstruction.
Inventors include S S Imambi; Kolukuluru Krishnamurty Raju; and Koneru Lakshmaiah Education Foundation.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: The present invention relates to a Multi-Domain Transformer-Based Framework for EEG Artifact Detection and Reconstruction for automatically identifying and reconstructing contaminated electroencephalography (EEG) signals. EEG recordings are frequently affected by ocular, muscular, motion, and other artifacts, which can reduce signal quality and adversely affect subsequent analysis. The proposed system receives contaminated EEG signals and performs preprocessing, segmentation, and normalization before extracting complementary temporal, frequency, and spatial features. A Transformer-based temporal module captures long-range dependencies in EEG signals, while frequency-domain processing extracts spectral characteristics and spatial learning models relationships among EEG electrode channels. The extracted features are integrated through a multi-domain feature fusion mechanism to generate a comprehensive representation of the EEG signal. An artifact detection module identifies contaminated regions and generates an artifact mask, which guides a reconstruction module in estimating and reconstructing the underlying EEG signal. The proposed framework is designed to reduce artifact contamination while preserving important neural characteristics of the original EEG recording. The reconstructed signals may be evaluated using signal-quality metrics including RMSE, SNR, PSNR, correlation coefficient, SAR, PNR, and RRE. The invention provides an automated and integrated approach for EEG artifact detection and reconstruction and may be applied in neurological disorder analysis, EEG classification, brain-computer interfaces, clinical research, and other EEG.
Disclaimer: Curated by HT Syndication.
