タイトル無し
スポンサーリンク
概要
- 論文の詳細を見る
In this paper, a computer recognition technique of sleep EEG patterns is proposed based on the statistical pattern recognition method for the purpose of automatic decision of human sleep stages.<BR>The sleep EEG has peculiar characteristics corresponding to each sleep stage and is grouped roughly into the following three patterns : <BR>i) High voltage exceeding 75μV and slow waves (δ-waves), <BR>ii) Alpha and/or low voltage waves mixed frequency activity, iii) Special waves such as sleep spindle, K-complex, spike, sharp-waves, saw-tooth waves, etc.<BR>In our method, two kinds of modeling and the associated feature extractions are considered for the EEG patterns i) and ii). One is a sinusoidal superposition modeling and pre-specified dominant frequency components are estimated as feature parameters by Kalman filtering. The other is a spline function modeling and parameters of the wave-shape such as smoothness, period, peak-to-peak, etc. are calculated by using the spline interpolated function.<BR>These processings in real time are performed for each few seconds of observation and the feature extraction results are summarized by using proper amount of observed training data in order to construct each pattern feature space.<BR>The testing data with pattern i) or ii) are classified by the existing statistical pattern recognition methods, e. g., Bayes decision rule and linear discriminant function scheme, based on the constructed feature spaces.<BR>It was verified by experimental results that the decision accuracy of our method is high (about 90%), and by extending the objects for processing to the special waves, this recognition technique may be applicable effectively to the on-line discrimination of EEG pattern variations.
- 一般社団法人 日本生体医工学会の論文
一般社団法人 日本生体医工学会 | 論文
- Evaluating the Workload Reduction of Automatic Vital Data Transmission
- Mental Fatigue Measurement Based on the Changes in Flicker Perception Threshold using Consumer Mobile Devices
- A Theoretical Study on a Computational Algorithm for Human Posture Estimation Based on Motion Capture of a Small Number of Markers
- Latest microscopic technique. From the principle to application. Development of three-dimensional microscopic system for operation of mini-pig fertilized eggs.
- A Study of the Automation of Cytodiagnosis (The Second Report):Cytophotometry Using Photomicroscanner and a Trial of Data-processing