期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2016
卷号：88
期号：2
出版社：Journal of Theoretical and Applied
摘要：Human actions are uncountable and diverse in nature; each action has its own characteristics and nature. Moreover, actions are sometimes different and sometimes very similar. Thus, it is rather challenging to implement a system that is capable of recognizing all human actions. However, the problem of recognition can be made simpler if the system of recognition is built gradually. Firstly, a set of required actions must be selected and then each action must be studied and analyzed to determine the most distinctive features that remain similar if different subjects perform the same action. The concept of action-based features is proposed and validated in this article. The system still has the ability to be extended to recognize more actions, simply by including contextual features of any added action. Experimental results have shown an outperforming performance with 100% accuracy based on the evaluation of UTKinect Action Data Set.
