计算机应用和软件说明.docVIP

第22卷第10期 计算机应用与软件 Vol.22, No.10 2005年10月 Computer Applications and Software Oct. 2005  HYPERLINK /pzqd/ 电脑配置清单  HYPERLINK /zzdnjc/  组装电脑教程  HYPERLINK /pzqd/2000y/  2000元电脑配置单 第 卷第 期 计算机应用与软件 Vol. No. 2005年1月 Computer Applications and Software Jan. 2005 基于Kinect传感器和Adaboost算法的人体动作识别研究 基于Kinect深度传感器提取人体骨骼信息，提出了一种有效的动作识别方法。该方法首先利用Kinect传感器获取人体骨骼三维位置信息和骨骼点的旋转信息，其中骨骼点的旋转信息用四元数的形式来表示；然后由获得的人体骨骼旋转数据构造候选特征曲线；最后用Adaboost算法对样本进行训练，得到相应的动作特征集和强分类器，并对强分类器进行融合，提出了动作识别力度来提高动作的识别率。实验结果证明，该方法对不同体型的人的动作都具有很高的识别准确率和鲁棒性。 Kinect深度传感器 Adaboost算法 四元数 候选特征 动作识别力度 中图分类号 TP391.9　　　　文献标识码 A 　　　　 Research on human action recognition based on Kinect sensor and Adaboost algorithm Zhu Xi’an1 Guo Dandan2 [作者英文名或拼音] 1,2(College of Information and Telecommunication Engineering, Beijing Information Science and Technology University, Beijing 100101) An effective Human action recognition method based on the human?skeletal?information?which is?extracted by Kinect depth sensor?is proposed in this paper. First, human bone three-dimensional position information and the rotation information of skeletal point expressed in the form of quaternion are acquired by?Kinect depth sensor. Then the candidate characteristic curves were constructed by human bone rotation data information. Finally, a corresponding set of operation characteristics and the strong classifier are obtained by training samples based on Adaboost algorithm, and the strong classifiers are fused. The action recognition  HYPERLINK /dict_result.aspx?searchword=%e5%8a%9b%e5%ba%a6tjType=sentencestyle=t=strength strength is proposed to improve the recognition rate movements. The experimental results show that the method for people of different body movements has a high identification accuracy and robustness. Kinect depth sensor Adaboost algorithm HYPERLINK javascript