When the robot recognizes a face in its field of view, it first grabs a snapshot of this face and extracts the relevant features and contours characterizing it. The robot then grabs a pen and start drawing the user's portrait, starting from the rough contours and adding details iteratively.