A New Marker-Less 3D Kinect-Based System for Facial Anthropometric Measurements

Several research fields like forensic and orthodontics, use facial anthropometric distances between a set of pairs of facial standard landmarks. There are tens of attempts to automatize the measurement process using 2D and 3D approaches. However, they still suffer of three main drawbacks: human manual intervention to fix the facial landmarks, physical markers placed on the face and required time for the measurement process. In this paper, we propose a new marker-less system for automatic facial anthropometric measurements based on Microsoft Kinect and FaceTracker API. This new approach overcomes the three measurement drawbacks. We statistically validated the system with respect to the caliper-based manual system, through experimental measurements and one-way ANOVA test comparisons on 36 subjects. The achieved successful percentage in the comparison is equal to 54,5 %.