A Research on Face Profile Surface Acquisition and Robot Trajectory Planning

Abstract

This study researches a method of the facial profile surface acquisition and robot trajectory planning on the human face. A Kinect 2 is firstly used to obtain the facial contour information. Then a deformation method based on 68 feature points is proposed to improve the accuracy of the facial model. Furthermore, a refinement and optimization of facial mesh model is constructed for robot trajectory planning and posture adjustment in working process. For complex facial contours and features, a cross-section method is suggested to plan discrete points of the facial operation trajectory, in which the robot posture can be controlled to changing with the trajectory normal curvature. Simulation and experiment demonstrates that the studied method is effective in 3D face model construction and robot movement control of the facial operation.