Screen Control System Based on Fingertip Tracking : AirFlick

Abstract: This paper suggests a method of extracting fingertips from 3D information obtained from single-depth cameras in a smart-device environment, and for controlling the screens without directly touching them. This method extracts fingertips from hand images, and controls the screen by reading the fingertips’tracking information. First, we extract hand areas in real time by utilizing depth information, and remove the noise by preprocessing the extracted hand areas to obtain the hand area through labeling. Next, we extract prospective fingertips from the areas and obtain final fingertips after a verification process. Fingertip movement information is indicated as a graph, and movement at a pace exceeding a certain critical level generates flick events. Finally, the key codes obtained from such events are relayed to application contents to control the screen.