Sentiment Analysis for Depression using Human Expression

Millions of people across the globe suffer from the debilitating mental illness known as depression. Early intervention and treatment of those suffering from depression is of the utmost importance in ensuring that the condition does not worsen. This study demonstrates a machine learning system for recognizing depression in an individual through their vocal characteristics and facial expressions. Psychological questions are posed, and their answers are recorded along with their voice and facial features, to evaluate their current emotional state.To develop this system, we employ AngularJS for the front end, Flask for the back end and Python for the machine learning process. Specifically, we utilize Python packages such as io, cv2, pickle, numpy, pandas, talk\_recognition, moviepy, pydub, keras, tempfile, flask, vander Sentiment and twitter_analysis. The data collected is then processed through video, audio and question-and-answer methods, and the results show that our system is able to detect mood disorder with 89\% precision, 92% accuracy, 87% recall, and 89% F1 outcomes. The conclusion that can be drawn is that these techniques have the potential to be effective tools in the early identification of depression.