Publication

Abstract : Brain Computer Interface (BCI) makes it possible for a machine and human brain to communicate directly. It serves as a valuable tool for individuals facing speech disabilities, physical limitations, and cognitive disorders. BCI operates by capturing signals from the brain known as electroencephalography (EEG) waves. In this paper, an AI-based assistive system that utilizes Brain-Computer Interface (BCI) technology is proposed. A non-invasive electrode band containing six electrodes is placed on the scalp. The brain signals are captured using electrodes and converted into raw EEG signals using Arduino. Filtering techniques like high-pass, low-pass, and notch filters are applied to improve signal quality by removing noise and artifacts. Power Spectral Density (PSD) analysis identifies usable frequencies. Features extracted include frequency-domain and time-frequency domain features. Transformer module convert EEG signals into characteristics using techniques like Short-Time Fourier Transform (STFT), teaching patterns to a Transformer-based model for anomaly detection. A translation algorithm facilitates communication between users speaking different languages. Translated text is displayed on user devices for comprehension.