Recognition of Eye Movement Electrooculogram Signals Using Dynamic Neural Networks

Human Computer Interaction provides a digital communication between the human and the physical world. This paper concentrates on tracking the eye movements through Electrooculography for HCI with the help of Neural Networks. Two feature extraction algorithms are used to extract the features from raw EOG signals for sixteen eye movements. The signals are classified into sixteen states using two networks namely Feed Forward Neural Network and Elman Neural Network. The performance of the proposed algorithms have an average classification efficiency of 83.36% and 98.50% for Singular Value Decomposition features and 84.60% and 98.46% for band power features using Feed Forward Neural Network and Elman Neural Network respectively. From the results it is observed that Elman Neural Network classifier using band power features outperforms the Feed Forward Neural Network classifier marginally.