EEG Based Automated Detection of Six Different Eye Movement Conditions for Implementation in Personal Assistive Application

Abstract

Contemporary human-machine-interfaces (HMIs) employ a wide range of human expressions to provide assistive support to the elderly and disabled population. Based on the disability type, expressions conveyed in terms of eye movements are often found to provide the most e cient way of communication. Nowadays, standard Electroencephalogram (EEG) based arrangements, used to analyze neurological states are also being adopted for the detection of eye movements. Although, a majority of the EEG-based state-of-the-arts researches either detects eye-movements in a lesser direction or uses a higher feature dimension with limited classi cation accuracy. In this study, a robust, simple and automated algorithm is proposed that uses the analysis of the EEG signal to classify six different types of eye movement. The algorithm uses discrete wavelet transform (DWT) on the EEG signals acquired from six different leads to eliminate a wide range of noise and artefacts. Then, two features per lead are extracted from the reconstructed wavelet coe cients and combined to form a binary feature map. Finally, a unique feature obtained from the calculated weighted sum of the binary map is used to classify six types of eye movements via a threshold-based technique. The algorithm presents high average accuracy (Acc), sensitivity (Se), speci city (Sp) of 95.85%, 95.83% and 95.83% respectively, using a single feature value only. Compared to other state-of-the-art methods, the adopted simple methodologies and the obtained results indicate the immense potential of the proposed algorithm to be implemented in personal assistive applications.

Mental stress detection using multimodal characterization of PPG signal for personal healthcare applications

“…E ciency of the chosen features is also evaluated using other standard classi cation techniques and the performance is calculated via four fold cross validation technique. For each of the mentioned classi er the overall classi cation performance is presented in terms of state-of-the-art statistical parameter as de ned in [27].…”