Wavelet domain extraction of features from single channel sleep EEG

Abstract

Sleep is just as important as diet and exercise. Humans spend about a third of their lives asleep. In the large data sets like Sleep Electroencephalogram (Sleep EEG), to do analysis it becomes tedious and time taken. Instead of considering the whole data, considering a few critical features from the signal makes the analysis simpler and the memory requirements are also less, since the analysis could be carried out on digital platform. A feature is a distinguishable sectional property obtained from a portion. Feature extraction depicts the number of feature to be extracted from the signal. Thus the feature extraction plays a pivotal role in the analysis of Sleep EEG. In this work we discussed the decomposition of Sleep EEG signal into required frequency bands and adopted feature extraction techniques of wavelet decomposition method to extract features from Sleep EEG signal by considering single channel EEG.