Real-time gesture recognition using 3D depth camera

Abstract

This paper proposes a vision-based human-computer interface system that can interpret a user's gestures in real time to manipulate windows and games. The system uses a 3D depth camera to capture depth data up to 30 frames per second. It extracts hand and fingertips from the acquired depth image. It recognizes the movement and click gesture by analyzing the location and shape change of fingertips. The proposed system is applied to play game “Angry Birds” instead of mouse and keyboard. Experimental results show that the proposed system is more robust than the method using a general camera.