本文提出了一种新的基于视觉技术的手指屏幕交互方法,利用手指在屏幕上的移动和点击等动作完成与计算机的交互.该方法使用两个摄像头从不同视角实时拍摄计算机屏幕,结合人手肤色和轮廓信息对手指进行实时跟踪,并根据两幅图像中的指尖位置和摄像头参数来判断指尖是否接触屏幕.实验结果表明,该方法可以完成鲁棒的手指跟踪和准确的点击检测,对光照条件没有严格的限制,实现了自然友好的人机交互操作,可以运用于智能交互、数字娱乐、穿戴计算等多种应用场合.
This paper presents a novel vision-based method for finger-screen interaction which achieves natural interactions through moving and clicking a single finger on a computer screen.The method uses two cameras at different viewpoints to capture video images covering the computer screen in real time.We integrate two cues of skin color and contour information to track finger' s motions in the framework of particle filter.The clicking actions on the screen surface are detected with the obtained fingertip positions and the geometry parameters of cameras.Experimental results demonstrate that the interactive method can realize robust tracking of finger motion,against complex background and varying illumination conditions,as well as the instant detection of finger clicking actions.This method achieves robust and natural human-computer interaction,and can be widely used in many applications,such as intelligent interaction,digital entertainment and wearable computing environment.