Luca Stocchi Master Thesis

The recent availability of the KinectTM, a low resolution 3D motion controller, offered interesting opportunities in the field of the facial expressions recognition. Although no studies have yet been conducted to evaluate both the reliability and the validity of this device, with this research we are interested to examine these two parameters by performing facial expression recognition tests in both real-time and offline mode. This paper discusses the application of two different approaches, the Emotional Facial Action Coding System (EMFACS), used to analyze data in real-time and the Local Zernick Moments (ZMs), used to detect expressions in offline mode. Experiments are performed using the FaceWarehouse and the VAP RGB-D databases with an overall recognition rate of 87.68%. Keywords—facial expression recognition, Microsoft Kinect, FACS, Zernike Moments.