作者: De-xin Wang , Zhi-hui Xiong , Mao-jun Zhang
DOI: 10.1007/S11042-011-0730-4
关键词: Computer science 、 Feature (machine learning) 、 Gesture recognition 、 Multi-touch 、 Surface computing 、 Gesture 、 Computer vision 、 Artificial intelligence 、 Image (mathematics)
摘要: In order to customize multi-touch gestures for different applications, and facilitate gesture recognition, an application oriented shape feature based description recognition method is proposed. this method, are classified into two categories, namely atomic combined gesture, where a combination of using temporal, spatial logical relationships. For description, users' motions mapped gestures, then semantic constraints extracted build the accessible relationships between entity states. trajectories projected onto image, every trajectory each other match with templates. Experiments show that independent platforms, robust manipulating differences users, it scalable reusable users applications.