基于改进凸分解的手势识别

Hand Gesture Recognition Based on an Improved Convex Shape Decomposition

提出了一种新的基于Kinect的实时静态手势识别方法,主要贡献包括:提出了一种简易可行的、结合图像深度信息与肤色信息的手势区域检测与分割方法;提出了一种改进的凸分解算法,对手势区域进行近似凸形状分解,以得到表征手势特征的骨架信息;采用基于路径相似性的骨架图匹配算法对手势进行匹配以实现识别.针对特定手势集进行了对比实验,实验结果表明,本文方法在识别结果的准确率以及算法的效率上都有着良好的表现.

Hand gesture recognition has become one of the most natural way of human communication with computer, However,the recognition accuracy and efficiency of the traditional methods still need to be improved.In this paper,a novel method was proposed based on Kinect for static gesture recognition in real time.Firstly,a simple and feasible hand detection and segmentation method was put forward based on depth information and skin color information.Then,a improved convex shape decomposition method was developed to obtain hand skeleton,and a gesture recognition method was put forward based on comparing the geodesic paths between skeleton endpoints.Finally,a comparative experiment was carried out for a specific gesture set.The experimental results show that,this method can provided a good performance in recognition accuracy and efficiency of the algorithm.