On the performance of wavelet transform improving Eye blink detections for BCI

Detection and determination of Eye blinks from Electro-occulography Signals are very helpful in Brain Computer Interfacing (BCI) Systems such as wheel chair controlling, cursor controlling and home automations. Eye ball movements are vital signs in some of the neurological disorders and represents the electrical activity of the muscles which steering the eye movements. EOG is an obtrusive, inexpensive and non-invasive means of recording eye ball movements. The source for EOG signal is cornea-retinal potential (CRP) and is generated due to the movements of eye balls within the conductive environment of the skull. While recording the EOG signal, it will be contaminated by electromyography (EMG) signal. As the EOG is a non stationary signal, the multi resolution analysis using wavelet decomposition offers the best solution to denoise and feature extraction of EOG signals. In this paper, proposed wavelet based method to detect eye ball moments from signal conditioned EOG. Eye links are detected with high sensitivity and specificity. Comparative wavelet analysis is performed by considering different statistical measures.