Wavelet based Novel Technique for Signal Conditioning of Electro-Oculogram Signals

In this paper, we present a novel and simple technique for signal conditioning of EOG signals which primarily involves denoising corrupted signals and post-processing for signal enhancement. Researches in the past have mainly focused on the EOG signals where the problem of removal of ocular artifacts from the electroencephalogram was dealt. We present, from a new perspective, a scheme which essentially deals with enhancement of EOG Signals. The non-stationary and time-varying EOG signals are processed using methodologies anchored on multiresolution analyses and the wavelet transform theory. Coiflet wavelets are used for subsequent removal of noise from the (awgn) corrupted EOG signals using the concept of coefficient thresholding. SURE is used for threshold selection. Its performance, in terms of SNR, is compared with strategies suggested by Birge-Massart and Donoho and Johnstone. Haar based wavelets of higher orders are used for post-processing of EOG signals. A pronounced advantage of post-processing of signals is that it facilitates the estimation of time instants and durations of intentional eye gestures which mainly find application in the development of human-computer interface based devices