Features and Recognition Algorithms in Sleep Staging and Drowsiness Detection by Using EEG

Abstract

Sleep quality is a significant problem worldwide, and drowsiness detection is meaningful to sleep staging as it can prevent many accidents with significant loss. Among all the methods, the method that combining the neuron science and the computer science to analyze the EEG signals has the best accuracy. The general steps of sleep staging include gathering the EEG signals, extracting the statistical features, putting the features into the model through recognition algorithms, and evaluating the model. The extraction of features and recognition algorithms are the most important parts of all the steps, which means the two steps can decide the precision of the method. This paper states and compares the different features and recognition algorithms used in this area and introduces some research using the related features and recognition algorithms. Thus, this paper can provide researchers valuable references of the feature and recognition algorithms and therefore develop more useful method for sleep staging in the future.