文献类型：会议论文

英文题名：Real-time forearm detection and tracking for HCl

作者：Zeng, Chengbin

第一作者：曾成斌

通信作者：Zeng, Chengbin|[14440d97b00e1ca4e9145]曾成斌;

机构：[1] Guizhou Institute of Technology, China

第一机构：贵州理工学院

通信机构：|贵州理工学院大数据学院

会议论文集：Proceedings of 2017 9th International Conference on Machine Learning and Computing, ICMLC 2017

会议日期：February 24, 2017 - February 26, 2017

会议地点：Singapore, Singapore

语种：英文

外文关键词：Cameras - Computer games - Graphic methods - Machine learning - Monte Carlo methods - Real time control - Stereo image processing

年份：2017

摘要：Natural human-computer interaction, as an active research topic in the area of computer vision, has widespread applications. In this paper, we propose a new method using forearm gesture as the input of the Tetris game via an ordinary camera. We first use histograms of oriented gradients and the skin color to detect the player's forearm. Then, the movement of the detected forearm is tracked using particle filter to obtain the directional information. Finally, the directional information is mapped into the input of the Tetris game. Compared to the other human-computer interaction (HCI) systems which use expensive equipments such as stereo cameras, our system only needs an ordinary camera. Experimental results show our method can efficiently detect and track the human forearm, and implement the real-time control for the Tetris game. ? 2017 ACM.