Figure 6.

(a) The model architecture of the CNN and the hybrid CNN-RNN models. The model selector block is a hypothetical block to show the difference between the two models: state 1 shows the CNN model and state 2 shows the CNN-LSTM model; (b) details of the inner layers in the CNN blocks; (c) details of the LSTM block. CNN: convolutional neural network, HRV: heart rate variability, PSD: power spectral density, RR: respiration rate, LSTM: long short-term memory, FC: fully connected, BN: batch normalization, and ReLU: rectified linear unit.