When the shoulder comes out of rest position or there are failures in face detection, the system loses its main point of reference. Thus,
the system no longer act as a detector, where each frame contained enough information to recreate all checkpoints and independent worked
of their predecessors. From this point forward, the system begins to act as a tracking, using back references to the generation of new data. The tracking functions require as input parameters an image and a region of interest. They identify each frame based on the previous frame, the new region of interest. These data are used as feedback to the face detector system. The points to be tracked are the shoulder arm and face, to either continue the three-dimensional drawing of the arm, avoiding a break in the immersion of the patient, as well as feedback to the system itself, accelerating the search function of the face.