Fig. 3.
Bidirectional LSTM model for sleep detection. The input xt is the feature matrix at time t, and the output yt ∈ [0, 1] is the estimated sleep probability. Output dimensions are denoted in each box.
Bidirectional LSTM model for sleep detection. The input xt is the feature matrix at time t, and the output yt ∈ [0, 1] is the estimated sleep probability. Output dimensions are denoted in each box.