Ocular Artifact Removal from EEG Data Using FCIF and FCFBCSP Algorithm with Modified DNN

This research offers two innovative techniques to improve electroencephalography (EEG)-based brain-computer interfaces (BCIs). The first technique, called Four-Class Iterative Filtering (FCIF), efficiently removes visual distortions from EEG data by using an iterative filtering procedure and statistical analysis. Using a Four-Class Filter Bank Common Spatial Pattern (FC-FBCSP) in conjunction with a Modified Deep Neural Network (DNN) classifier, the second method aims to improve motor imagery classification accuracy by identifying unique spatial patterns linked to different tasks. The work illustrates the better performance of FCIF and FC-FBCSP over baseline techniques with a comprehensive experimental validation. These novel techniques have excellent for improving the efficiency and reliability of BCIs, especially for tasks involving the classification of motor imagery. The research’s conclusions not only develop BCI technologies but also open up new channels for real-world applications in contexts that prioritize users. In conclusion, this research offers significant approaches, FCIF and FC-FBCSP, that tackle important issues with EEG-based BCIs and offer a way forward for more dependable and effective BCI systems. The research’s conclusions have important implications for the creation of BCI applications that are both realistic and easy to use.