数字家庭环境中双手交互技术研究  

Research on Double Hand Interaction Technology in Digital Home Environment

作　　者：杨扬[1] 王志良[1] 杨溢[1] 

机构地区：[1]北京科技大学信息工程学院,北京100083

出　　处：《计算机工程》2011年第4期29-30,42,共3页Computer Engineering

基　　金：国家"863"计划基金资助项目(2007AA01Z160)

摘　　要：采用基于视觉的双手手势控制方法,构建数字家庭环境。在获取肤色图像后,通过面积过滤和图像方位分辨出左右手的位置,使用形状过滤算法识别出指尖的位置。分析左手静态手势判断所控制的家电,同时跟踪右手指尖移动的路径,通过方向离散化得到方向向量,使用Baum-Welch算法训练方向向量,用Viterbi算法识别动态手势,实现家电状态的控制。实验结果表明,该方法能有效地在数字家庭环境中完成普通家电的控制功能。Using vision-based hand gesture control method, this paper constructs a digital home environment. After skin color binary image is obtained, area filtering and image orientation difference are adopted to identify left and right hand position. Shape filter algorithm is proposed to track the location of fingertips. Static gestures of left hand are analyzed to control the home appliances, and the path of right hand fingertip movement is tracked, discrete direction vectors are calculated. Baum-Welch algorithm is used to train the discrete direction vector, Viterbi algorithm for dynamic gesture recognition is used to achieve home appliances control. Experimental result shows that the method can effectively complete control function of home appliances in digital home environment constructed.

关 键 词：手势识别 数字家庭 双手交互 隐马尔科夫模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]