Decision Tree Algorithm for Depression Diagnosis from Facial Images

Depression refers to a condition in which mental function continues to deteriorate and adversely affects daily life. Therefore, it is very important to diagnose depression accurately in the early stages. This study extracted depression-related features such as head pose, gaze, and facial expression changes from facial images and proposed a model for diagnosing depression using a decision tree. The proposed model learned the data using two databases. The performance of the algorithm was evaluated through the confusion matrix, receiver operating characteristic (ROC) curve, and achieved value 0.99 for area under the curve (AUC).