基于Leap Motion的手势识别在虚拟沙画中的应用研究

发布时间：2018-06-05 21:25

本文选题：Leap + Motion　；参考：《中北大学》2017年硕士论文

【摘要】：随着虚拟现实技术的快速发展和体感交互设备的不断推广,人们越来越倾向于使用自然、方便的方式实现人机交互,尤其是手势交互。因此,虚拟现实系统中手势识别的研究与发展具有重要的意义。沙画创作是21世纪初期新兴的一种艺术创作方式,沙子动画搭配流畅跌宕的背景音乐给人们带来全新的视听感受。近些年,虚拟现实技术逐渐融入沙画艺术中,形成虚拟沙画。因此,本课题以虚拟沙画为应用背景,对手势识别算法理论及Leap Motion的关键技术进行了系统的研究,改进手势识别算法,设计并实现了基于Leap Motion和Unity 3D的虚拟沙画系统。主要工作内容如下:(1)深入研究手势识别算法,结合稀疏表示方法和压缩感知理论,引入基于稀疏表示的分类(SRC)算法;在此基础上,结合K-最近邻(K-NN)算法,对SRC算法进行改进,提出KNN-SRC算法,并从理论和实验两方面验证其可行性。(2)结合KNN-SRC手势识别算法,设计并实现了基于Leap Motion的手势识别模型。本文充分利用Leap Motion精确捕获人手信息的特性,完成数据的采集、特征向量的提取、指尖位置的检测及跟踪。同时,改进特征提取算法,优化跟踪算法,最终通过实验对相关改进内容的有效性进行测试,为实现虚拟沙画系统奠定基础。(3)将构建的基于Leap Motion的手势识别模型及相关改进算法应用到Unity 3D平台搭建的虚拟沙画系统中。编写脚本文件定义手势语义并在手势识别后实现沙画绘制,采用粒子系统模拟落沙效果,并导入在3Ds Max软件中构建的虚拟手模型提升系统的真实感。
[Abstract]:With the rapid development of virtual reality technology and the promotion of interactive devices, people are more and more inclined to use natural and convenient ways to realize human-computer interaction, especially gesture interaction. Therefore, the research and development of gesture recognition in virtual reality system is of great significance. Sand painting is a new art creation method in the early 21st century. Sand animation with smooth and volatile background music brings people a new audio-visual feeling. In recent years, virtual reality technology gradually into the art of sand painting, forming virtual sand painting. Therefore, based on the application background of virtual sand painting, this paper systematically studies the theory of gesture recognition algorithm and the key technology of Leap Motion, improves the algorithm of gesture recognition, and designs and implements a virtual sand painting system based on Leap Motion and Unity 3D. The main work of this paper is as follows: (1) We deeply study the gesture recognition algorithm, introduce the sparse representation classification algorithm based on sparse representation and compressed sensing theory, and then improve the SRC algorithm with the K-NNN algorithm, which is based on the K-NNN algorithm. The KNN-SRC algorithm is proposed, and the feasibility of the algorithm is verified theoretically and experimentally. Combining with the KNN-SRC gesture recognition algorithm, a gesture recognition model based on Leap Motion is designed and implemented. In this paper, we make full use of the characteristics of Leap Motion to capture human hand information accurately, complete data acquisition, feature vector extraction, fingertip location detection and tracking. At the same time, we improve the feature extraction algorithm, optimize the tracking algorithm, and finally test the effectiveness of the related improved content through experiments. In order to realize the virtual sand painting system, the gesture recognition model based on Leap Motion and the related improved algorithm are applied to the virtual sand painting system built on the Unity 3D platform. The script file defines gesture semantics and realizes sand drawing after gesture recognition. Particle system is used to simulate the effect of sand drop and the virtual hand model built in 3Ds Max software is used to improve the reality of the system.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP391.9

【参考文献】