Optimizing Residual Networks and VGG for Classification of EEG Signals: Identifying Ideal Channels for Emotion Recognition

Table 2

Recent research on SEED dataset.

Classifier algorithm/year	Data input	Accuracy (%)
Dynamic graph CNN [19]/2018	Differential entropy (DE)	79.95
Logistic regression classifier [20]/2018	DE	72.47
GRSLR (graph regularized sparse linear regression) [21]/2018	DE, Hjorth features	88.41
Bidirectional LSTM [22]/2019	DE/Power spectral density (PSD)	94.96/86.27
Graph convolutional broad network (GCBN) [23]/2019	DE	94.24
CNN + LSTM [24]/2019	DE	89.88
Variational pathway reasoning (VPR) [25]/2019	DE	94.3
Sequential backward selection SVM [26]/2019	Hjorth features, standard deviation, sampling entropy, wavelet entropy	89
Spiking NN [27]/2020	DWT, FFT, variance	96.67