Table 1.

Re-training computational times

	Re-training time (s)	Detection time (s)
XGBoost	36.5	0.8
SVM	14.9	4.9
LSTM	130.6	37.4
2D-CNN	29.8	12.9
1D-CNN	24.4	10.7

The first column shows how long the re-training lasted with each model for 5 epochs, using a training batch of 32 samples (in seconds). The second column shows how long the re-trained models took to process 1 h 55 min of session (in seconds).