节点文献
连续形变图像的追踪识别算法及其在手势识别系统中的应用研究
Research on Tracking Recognition Algorithms of Continuous Deformational Image and Its Application in Hand Gesture Recognition
【摘要】 提出了一种连续形变图像的追踪识别算法,通过引入网格等基本概念,建立了一套基本理论,通过揭示网格与连续形变之间的内在联系,避免了传统方法中复杂的拓扑分析.该算法把追踪与识别过程有机地统一起来,既具有一定的理论基础,又可以灵活地满足工程中不同的精度需要.最后讨论了该算法在手势识别系统中的应用.实验表明,算法取得了令人满意的灵活性、适应性和时间复杂度.
【Abstract】 The tracking recognition of continuous deformational images is one of the fundamental and core problems in robot visual systems.A new algorithm to track continuous images is put forward in the paper.By means of introducing some new basic concepts upon which the fundamental theories are built,exposing the internal relationships between grid and continuous deformation,traditional complicated topological analyses are avoids.As a result,the algorithm developed in this paper is credible and flexible,and satisfied different needs with different precisions.At last,the application in hand gesture recognition is discussed.Experimental results show that this algorithm is good at flexibility,adaptabilities and time complexity.
【Key words】 continuous deformational; topological structure; tracking recognition; human-computer interaction(HCI); human hand tracking;
- 【文献出处】 计算机学报 ,Chinese Journal of Computers , 编辑部邮箱 ,2006年01期
- 【分类号】TP391.41
- 【被引频次】6
- 【下载频次】505