Fig. 2: Overview of developed model for objective evaluation of mental health status.
From: Identifying mental health status using deep neural network trained by visual metrics
Three different networks with the same architecture are trained to evaluate HHI, STAI, and WEMWBS metrics. The HHI, STAI, and WEMWBS are metrics used for subjective evaluation of hope, stress/anxiety, and mental well-being, respectively. Blocked data through visual metrics time series were used as input to the CNN-LSTM model. In this study, sliding window length ‘L’ is 51 and 20 visual metrics are considered as dimension of input to the network. The output of the model is level of mental well-being (evaluated by WEMWBS metric), anxiety/stress (evaluated by STAI metric and hope (evaluated by HHI metric) based on three categories of low (class label: 0), intermediate (class label: 1), and high (class label: 2) categories.
