DWT-Net: Seizure Detection System with Structured EEG Montage and Multiple Feature Extractor in Convolution Neural Network

Algorithm 1. Algorithm for state machine post-processor

Require:Pseizure, Pnormal, Pup

Output:Result_n, n =1,2,…

1: Initialize System State State ⇐ Negative

2: while New EEG window do

3:  ifState == Negativethen

4:   ifPseizure < Pnormalthen

5:    next State ⇐ Negative

6:    Result_n ⇐ Normal

7:   else

8:    next State ⇐ Positive

9:    Result_n ⇐ Epileptic

10:   end if

11:  end if

12:  ifState == Positivethen

13:   ifPseizure + Pup < Pnormalthen

14:    next State ⇐ Smooth

15:    Result_n ⇐ Normal

16:   else

17:    next State ⇐ Positive

18:    Result_n ⇐ Epileptic

19:   end if

20:  end if

21:  ifState == Smooththen

22:   ifPseizure + Pup < Pnormalthen

23:    next State ⇐ Negative

24:    Result_n ⇐ Normal

25:    Result_n−2 ⇐ Normal

26:   else

27:    next State ⇐ Positive

28:    Result_n ⇐ Positive

29:    Result_n−1 ⇐ Positive

30:   end if

31:  end if

32: end while

Algorithm 1. Algorithm for state machine post-processor