Classification of Mental Stress Using CNN-LSTM Algorithms with Electrocardiogram Signals

Table 1

Classification layers used to evaluate stress signals using CNN-LSTM.

Number	Layer	Activation	Weights	Bias
1	Sequence input layer	124 × 124 × 3	—	—
2	Sequence folding layer	124 × 124 × 1	—	—
3	Convolution 2D layer	124 × 124 × 6	5 × 5 × 3 × 6	1 × 1 × 6
4	Batch normalization layer	124 × 124 × 6	—	—
5	Max pooling layer	62 × 62 × 6	—	—
6	Convolution 2D layer	62 × 62 × 12	3 × 3 × 6 × 12	1 × 1 × 12
7	Batch normalization layer	62 × 62 × 12	—	—
8	Max pooling layer	31 × 31 × 12	—	—
9	Sequence unfolding layer	31 × 31 × 12	—	—
10	Flatten layer	11532	—	—
11	LSTM layer	200	Input: 800 × 11532 recurrent: 800 × 200	800 × 1
12	Fully connected layer	2	2 × 200	2 × 1
13	Softmax layer	2	—	—
14	Classification	—	—	—