Analysing the Effective Psychological State of Students using Facial Features

Abstract

The psychological state of college students plays an important role in their academic performance. Negligence of one’s mental and emotional health can lead to severe problems such as anxiety, stress, depression, etc. Quick detection and treatment of these problems are essential for the mental health of students. Depression detection in college students is very tough, largely due to the sheer population of students, along with the fact that they may not be aware of their mental problems. Furthermore, because of the stigma surrounding mental health, students may be concealing their psychological issues from everyone. An automatic depression detection system that can predict if students are dealing with depression is required. The proposed system uses video footage of college students to extract the facial features by use of the KLT Algorithm (Face-Detection/Tracking) and Gabor Filters to examine them with an intent to identify symptoms of depression. The final step is to classify the students as having depression or not by utilizing SVM based classifier. The system has been trained with datasets having images of faces along with the classification of whether they are contempt, disgusted, or happy. The level of these emotions in the extracted frames is analyzed to detect depression.