摘要
随着虚拟现实技术的发展和手势识别技术的不断提升,人们对自然交互的方式不断提出迫切需求。对于交互体验的游戏而言,尤为突出。此前的研究主要利用普通摄像头采集图像,获取手势的运动形态及像素信息,无法获得具有深度信息手势的三维特征。随着LeapMotion等深度传感器的出现,更多的深度信息可以被获取,为识别复杂的手势提供了可靠的数据保障。文章使用一种基于深度经网络的游戏交互手势识别系统,与以往的方法相比,具有更好的性能。
With the development of virtual reality technology and the continuous improvement of gesture recognition technology, people continue to put forward an urgent need for the way of natural interaction. For interactive experience of the game, especially prominent. Previous studies mainly used ordinary cameras to collect images to obtain the motion form and pixel information of gestures, but could not obtain the three-dimensional features of gestures with depth information. With the emergence of depth sensors such as Leap Motion, more depth information can be obtained, which provides a reliable data guarantee for the recognition of complex gestures. In this paper, a game interactive gesture recognition system based on depth neural network is used, which has better performance than the previous methods.
出处
《科技创新与应用》
2019年第20期22-24,共3页
Technology Innovation and Application
基金
教育部“春晖计划”项目(编号:Z2017027)
教育部计划项目(编号:2017B00020)
关键词
虚拟现实
手势识别
卷积神经网络
virtual reality
gesture recognition
convolution neural network
