Identifying Facial and Body Movement Through CNN And Monocular Tracking Algorithm

Facial Expression movement could analyze the true meaning behind the person's intention, but people could lie and fabricate their real expression by restraining the facial muscle. However, multitasking and restraining, both facial and body movement could reveal their true intention. Facial and body movement studies have been conducted to study the true meaning behind someone's movement. Through these studies, we will subtract a few methods that are accurate to the true result through the testing that have been conducted. We proposed the utilization of Seetaface method and a hybrid of deep learning approach with inertial sensors to analyze human true emotions. Through the Seetaface method, we will analyze the person's facial movement, which consists of the important facial features such as eyes, nose and mouth. Subsequently, the hybrid of deep data and facial features will detect and analyze a person's body movement through a camera and record it to the body models data. After successfully processing both of the methods, our findings show that facial detection is capable of pointing to the facial feature and expression of important features, however hindered by face shifting and obstruction. In body detection, the true action matched 100% successfully with the prediction of the action result after further evaluation using hybrid best results.