An Intelligent System for Prediction of Depression based on Facial Emotions and Textual Data

Globally, a significant portion of the population experiences depression or stress due to a range of factors at various stages of life. In contemporary society, the fast-paced nature of daily life contributes to the gradual development of depression, as ongoing stress accumulates over time. Artificial Intelligence based systems contribute significantly to the healthcare systems by providing timely intervention and automating the decision-making process. Deep learning-based algorithms are capable of recognizing facial expressions and identifying as well as predicting an individual's mental state. In this research article we propose a depression detection system based on deep learning. In the proposed system, a combination of facial expression and text data are used to predict whether a person is depressed or not. We utilize Convolutional Neural Network (CNN) for the extraction of features for the detection and prediction for depression. The proposed model was tested on FER dataset along with a Patient Health Questionnaire also known as PHQ-9 for the psychological analysis. With experiments our system obtained an accuracy of 92% with FER dataset.