The columns in the table correspond respectively to the correlation between: (a) 8 T-F features vector ({FS 1, FS 2,..., FS 8}), and (b) 13 T-F features vector ({FS 1, FS 2,..., FS 8, FI 1,..., FI 5}).
The results between parentheses are the classification results using the ten signal-related features {FS 1, FS 2,..., FS 10}.
Proof Following the proof of Theorem 1, FS ≠ ∅.
The T-F feature set {FS 1, FS 2,..., FS 10, FI 1,..., FI 5} was extracted from the TFD of each EEG segment of length T seconds.
We notice that the classification results using the feature vector {FS 1, FS 2,..., FS 10, FI 1,..., FI 5} with the SVM-based classifier, are better compared with the results using the feature vector {FS 1, FS 2,..., FS 10}.
The total number of features for this study was therefore 15 for the adult EEG data ({FS 1, FS 2,..., FS 10, FI 1,..., FI 5}) and 13 for the newborn EEG data ({FS 1, FS 2,..., FS 8, FI 1,..., FI 5}).
Table 6 shows that the classification results using the combined T-F signal & image related feature vector {FS 1, FS 2,..., FS 8, FI 1,..., FI 5} are better compared to results using only the T-F signal-related feature vector {FS 1, FS 2,..., FS 8}.
The results between parentheses are the classification results using the 10 features {FS 1, FS 2,..., FS 10} (without the features based on T-F image processing techniques) and the multi-class SVM classifier [14 16].
A 1 fs simulation time step was used.
All patients had FS ≥1 and were divided in two groups; patients with given FS equal to 1 (FS = 1) and patients with FS equal to or more than 2 (FS ≥2).
In the layer 6 sample, 1 neuron was LS, 1 FS, and 3 intrinsically bursting.
