Hand Gesture Recognition for Human-computer Interaction Using Moment Invariants and Neural Network C

• 作 者: RanChen;ShuangyuanWang;LiangGong;ChengliangLiu;
• 机 构: School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, China
• 出 处: 上海交通大学第16期博士生学术论坛——机械百年·智能制造·智造中国学术论坛
• 发表日期: 2013年期

　　Gesture recognition is of utmost importance in intelligent human-computer interaction (HCI).In this paper, an algorithm using moment invariants and BP neural network classifier for hand gesture recognition is proposed.Seven invariant moments are extracted after the image segmentation, which quantitatively describes the targeted hand gesture, and then the trained BP neural networks are applied to recognize the gesture.The algorithm is validated and achieves a high recognition rate of 99% in experiments, which makes it feasible for practical HCI in some restricted environments.