基于Kinect动态手势识别的改进DTW算法研究与实现

Improvement and Implementation of DTW Algorithm on Dynamic Gesture Recognition based on Kinect

为了解决传统算法在动态手势识别中存在的准确率不高、难以适应复杂背景等问题,本文提出了一种基于Kinect传感器的改进DTW算法动态手势识别方法.使用Kinect Studio和Visual Gesture Builder训练建立手势模板库,对传统的DTW算法进行限制搜索范围和设置失真度阈值两点改进,利用改进后的DTW算法对提取出来的人物指定关节信息和模板信息进行匹配,从而识别出手势.实验结果表明,改进DTW算法具有较高的动态手势识别准确率,并且能够更好地适应复杂环境.

In order to solve the problems of low accuracy and difficult adaptation to complex background in traditional dynamic gesture recognition,this paper proposes an improved DTW algorithm based on Kinect sensor.It creates the Gesture template library by using Kinect Studio and Visual Gesture Builder,and improves the traditional DTW algorithm by limiting the search range and setting the distortion threshold.In order to recognize hand gestures,the improved DTW algorithm is used to match the joint angle information and the template information.The experimental results show that the improved DTW algorithm has high accuracy of dynamic gesture recognition and can better adapt to the complex environment.