Artifical Intelligance based Real Time Lie Detector Using Eye Gaze Pattern

Deception is a sophisticated talent that takes place in human social interactions. Lying may be described as the act of hiding the truth using a false statement. Traditional methods like polygraphs, which are used for lie detection, involve the monitoring of sweat, respiratory rate, pulse rate, and blood pressure. These methods are not only invasive but also expensive. However, it also requires a skilled human interviewer to perform the questioning and evaluate the data. Moreover, those techniques are not always reliable. The need for an improvised way of detecting the deceitfulness of a person has become critical in the society we live in. The lie detection here is based on the hypothesis that liars experience more cognitive demand than truth-tellers. Lie detection will be performed in real-time video by capturing the eye gaze pattern. The gaze can be either to the left or to the right, which is based on the truthfulness of the person in the video. The eye gaze pattern is analyzed with the help of OpenCV using the dlib library and a pre- trained network called shape_predictor_68_landmarks.