现代电子技术2017，Vol.40Issue(11)：61-64,4.DOI:10.16652/j.issn.1004-373x.2017.11.016
特征降维和高斯混合模型的体育运动动作识别
Sports action recognition based on feature dimension reduction and Gaussian mixture model
区峻 1石千惠1
作者信息
- 1. 广西师范大学 体育学院 体操教研室,广西 桂林 541004
- 折叠
摘要
Abstract
In order to overcome the shortcomings of the current sports action recognition methods,and obtain the optimal recognition effect of sports action,a new sports action recognition method based on feature dimension reduction and Gaussian mixture model is proposed. The video image of sports action is collected to extract the feature vector of the sports action. The ran-dom projection algorithm is used to reduce the dimension of the feature vector. The Gaussian mixture model is used to study the training samples after dimension reduction to construct the sports action recognition model. The performance of the method is tested with various sports action datasets. The results show that the proposed method can obtain the satisfied recognition result of sports action,and the recognition accuracy is higher than that of other sports action recognition methods.关键词
体育运动动作识别/随机投影/高斯混合模型/特征向量降维Key words
sports action recognition/random projection/Gaussian mixture model/feature vector dimension reduction分类
信息技术与安全科学引用本文复制引用
区峻,石千惠..特征降维和高斯混合模型的体育运动动作识别[J].现代电子技术,2017,40(11):61-64,4.
