적외선 센서를 이용한 손 자세 인식

This paper deals with a hand gesture recognition system using Leapmotion infrared (IR) sensor. The hand region is extracted from the IR image intensity. The proposed algorithm for hand gesture recognition first estimates the center of hand palm using the distance transform. Then, each finger is identified by approximated curvature and convexity in the hand region. The proposed algorithm recognizes various gestures of hand by combining the number of fingers with motion of fingers. The proposed algorithm is focused on the mobile implementation in real-time. This paper also implements some user interface and interactions to show the recognition performance in the AR/VR system. It is expected that the proposed algorithm for hand gesture recognition is useful in real-time mobile AR/VR systems.