下載/瀏覽Download

Download Report

Transcript 下載/瀏覽Download

Chairman:Hung-Chi Yang
Presenter:Yu-Kai Wang
Advisor:Yeou-Jiunn Chen
Date:2013.10.30
0






Introduction
Paper review
Purposes
Materials and Methods
Future works
References
1

Sleep
◦ Patterns and waves of EEG
 The two most important characteristics of EEG elements
 Frequency
 Amplitude

The frequency range is divided into four bands
◦
◦
◦
◦
Beta (12-30 Hz)
Alpha (8-12 Hz)
Theta (4-8 Hz)
Delta (0,1-4 Hz)
2


Effective diagnosis and treatment of patients with sleep
Our objective is to utilize an classifier Using
◦
◦
◦
◦

Energy
Entropy
Frequency band
GMM
Features extracted from EEG characteristic waves
◦ To develop an effective automatic sleep stage classification
system using only a single EEG channel
3

Data acquisition


The sleep recordings utilized are obtained from the
Sleep-EDF database ,from the PhysioBank
Eight full sleep recordings from Caucasian


Aged from 21 to 35
Were not on any medication at the time of the
data collection
4

Feature extraction

First used six FIR bandpass filters

There are a total of 3000 samples in each characteristic
wave in each 30s epoch
The sampling rate of EEG signals equals 100 Hz
EEG signals from each 30s epoch by using


N
Energy   X i2
(1)
i 1
5

Sample-entropy
◦ This is the rate of new information producted in a
dynamic system
◦ The negative natural logarithm of the conditional
probability
 Two sequences similar for m points would remain similar
at the next point
◦ A lower value of SaEn indicates
 More self-similarity in the time series
SampEn=-log A/B
(2)
A = no of template vector pairs having d[X m (i),X m (j)] < r of length m+1
B = no of template vector pairs having d[X m (i),X m (j)] < r of length m
6

Support Vector Machines (SVM)
◦ SVM is a supervised learning method
 Classification
 Regression
◦ Training of SVM is to find the optimal hyperplane
(thick solid line)
 Separates the samples from two classes (circles vs.
squares) with maximum margin
7
◦ To understand the essence of SVM classification, one
needs only to grasp four basic concepts




The separating hyperplane
The maximum-margin hyperplane
The soft margin
The kernel function
8

Neaural Network
◦ Using Matlab toolbox (nntool)
 Import Input data(train data)
 Target data(stage)
 Sample data(test data)
 Create newnetwork(Feed-forward backprop)
 Set training epochs(3000 epochs)
 Simulation
 Performance
Source:google image
9







[1] R. Agarwal, J. Gotman, Computer-assisted sleep staging,, IEEE Trans. Biomed. Eng. 48 (12)
(2001) 1412–1423.
[2] S. Aydin, H.M. Sarao˘glu, S. Kara, Singular spectrum analysis of sleep EEG in insomnia, J.
Med. Syst. 35 (4) (2011) 457–461.
[3] C. Berthomier, X. Drouot, M. Herman-Stoı¨ca, P. Berthomier, J. Prado, D. Bpkar- Thire, O.
Benoit, J. Mattout, M. d’Ortho, Automatic analysis of single-channel sleep EEG: validation in
healthy individuals, Sleep 30 (11) (2007) 1587–1595.
[4] A.G. Correa, E. Laciar, H.D. Patin˜ o, M.E. Valentinuzzi, An automatic sleep-stage classifier
using electroencephalographic signals, Int. J. Med. Sci. 1 (1) (2008) 13–21.
[5] S. Charbonnier, L. Zoubek, S. Lesecq, F. Chapotot, Self-evaluated automatic classifier as a
decision-support tool for sleep/wake staging, Comput. Biol. Med. 41 (6) (2011) 380–389.
[6] K.I. Funahashi, Y. Nakamura, Approximation of dynamical systems by con- tinuous time
recurrent neural networks,, Neural Networks 6 (6) (1993) 801–806.
[7] L.A. Feldkamp, G.V. Puskorius, A signal processing framework based on dynamic neural
networks with application to problems in adaptation, filtering, and classification, Proc. IEEE 86 (11)
(1998) 2259–2277.
10



[30] M.E.Tagluk,N.Sezgin,M.Akin,Estimationofsleepstagebyanartificial neyral
networkemployingEEG,EMG,andEOG,J.Med.Syst.34(4)(2010) 717–725.
[31] J.S.Wang,C.S.G.Lee,Self-adaptiveneuro-fuzzyinferencesystemsforclassi- fication
applications,,IEEETrans.FuzzySyst.10(6)(2002)790–802.
[32] L.Zoubek,S.Charbonnier,S.Lesecq,A.Buguet,F.Chapotot,Featureselection for
sleep/wakestagesclassificationusingdatadrivenmethods,,Biomed. Signal
Process.Control2(3)(2007)171–179.
11
Thank You
For Your Attention
12