Figure 2.

Structure of the fall risk level classification model with mixed input data. The temporal input datum (upper branch) is a moving window of 64 timestamps with the three axes of each sensor (accelerometer and gyroscope). This branch of the model comprises a series of convolutional and dropout layers and LSTM to automatically extract the temporal characteristics of the signals. The branch with the frequency information (lower branch) is the spectrogram image of the temporal signal. This branch comprises convolutional layers to extract information from the images. Acc.: accelerometer signal; Gyr.: Gyroscope signal; V: vertical; ML: medial-lateral; AP: anterior-posterior. This figure was self-made, created by Adobe InDesign^® 2023 (version 18.4).