摘要
针对运动手势的动作识别问题,提出了利用均值漂移算法的动态手势识别方法。首先,利用均值漂移算法跟踪捕获手势动作;然后,对样本手势数据进行滤波和处理以选择最佳手势;最后,利用SVM进行手势识别。实验结果表明,提出的方法在运动流方面的手势识别率达到80%以上,效果非常乐观。同时,随机挑选了一个测试集,与其他方法进行比较,也证实了提出方法的优越性。
For the problems of motion gesture recognition,a dynamic hand gestures recognition method using Corrected Background-weighted Histogram(CBWH)is proposed.Firstly,the mean shift algorithm is used to track capture gestures.Secondly,the sample gesture data is filtered and processed to select the best gesture.Finally,SVM is used to identify gestures.Experimental results show that the proposed method in terms of flow movement gesture recognition rate above 80%,the effect is very optimistic.At the same time,a test set is selected randomly and compared with other methods.The superiority of the proposed method is also proved.
作者
李荣
Li Rong(Internet of Things Engineering College,Taihu University of Wuxi,Wuxi Jiangsu 214064,China)
出处
《信息与电脑》
2017年第12期93-95,共3页
Information & Computer
关键词
动态手势识别
均值漂移
序列对比算法
人机交互
支持向量机
dynamic gesture recognition
CBWH
sequence comparison algorithm
human-computer interaction
SVM
