摘要
文章首先给出了基于角度的动力学模型及其特征值,并提出了基于SCG神经网络的静态特征值识别算法和基于模板匹配的动态特征值识别算法。使用该文提出的动态时间规整算法和手势分割算法建立的动态手势识别系统,实践证明具有较好的实时性和识别率。
In this paper the kinetic model of angle based hand gestures and its eigenvalue are presented,then the SCG Neural Network Algorithm for static recognition and the Template Matching Arithmetic for dynamic recognition are pro-posed.With the Dynamic Time Warping Arithmetic and the segmentation of hand gestures,a system of dynamic hand gestures' recognition is set up.Experiment results indicate that it is a real-time system with high recognition rate.
出处
《计算机工程与应用》
CSCD
北大核心
2002年第19期73-75,共3页
Computer Engineering and Applications
基金
国家部委预研基金项目
