Wavelet Transform and Circulant Singular Spectrum Analysis for Improved EOG Artifact Elimination from Electroencephalogram Signals

The ability to learn, focus, remember, and make decisions are all necessary for completing various jobs. One limitation of the human mind is its limited capability to process & integrating. The mental strain that these finite resources endure is known as cognitive load. Variables related to psychological wellness, such as worry, stress, or changes in respiration, heart rate variability, conductance of the skin, or blood pressure, might alter pattern of cognitive load. The four goals of this thesis are based on the assessed research gaps: Automated and effective artifact reduction from EEG signals is achieved. Secondly, the effectiveness of identifying cognitive strain using EEG signals. Thirdly, cognitive stress detection using single-lead EEG. The fourth step is to develop a method for detecting cognitive fatigue on an individual basis using deep learning. In order to separate EOG-polluted EEG signals into their essential method components, the CiSSA method was used. The different parts of the artifact signals were identified using kurtosis and energetic values. The suggested CiSSA-DWT approach achieves an accuracy of about 99.83% when it comes to removing artifacts while maintaining the crucial EEG signal quality. When compared to other existing approaches, CiSSA-DWT is the most precise method since it effectively removes artifacts.