MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641079794 A) filed by Karunya Institute Of Technology And Sciences on June 29, 2026, for System For Multi-Class Motor Imagery Eeg Signal Classification Using Spectrogram Analysis And Efficientnet-B0 Network.
Inventors include Dr. M. S. P. Subathra; Mr. S. R. Subin; Ms. Thejaswini Kishore; Mr. P. Silvanus Paul; Mr. Jp. Ragavendra Prasath; Dr. S. Thomas George; Dr Esther Daniel; and Ms. Swathi Sennappan.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: TITLE: SYSTEM FOR MULTI-CLASS MOTOR IMAGERY EEG SIGNAL CLASSIFICATION USING SPECTROGRAM ANALYSIS AND EFFICIENTNET-B0 NETWORK APPLICANT: KARUNYA INSTITUTE OF TECHNOLOGY AND SCIENCES ABSTRACT The present invention discloses a system for motor imagery electroencephalogram (EEG) signal classification and neurorehabilitation. The system includes an EEG signal acquisition assembly with Ag/AgCl electrodes, preprocessing circuitry to clean the signals, and a signal transformation processor to convert the signals into spectrogram images using Short Time Fourier Transform (STFT). A classification computing unit, comprising an embedded NVIDIA Jetson Nano board with CPU and GPU, executes a trained EfficientNet B0 deep learning model stored in memory with TensorFlow and PyTorch frameworks. The system further includes a control interface unit with a display device that provides real time feedback and translates the classified outputs into commands for external devices such as robotic limbs, exoskeletons, or virtual reality systems.
Disclaimer: Curated by HT Syndication.
