图形用户界面(GUIs)是当今人机交互技术(HCI)的主要平台,它使得普通人都会使用计算机.然而,为了适应更广泛的应用,人机交互技术需要向自然、直观、自适应的交互界面发展,而感知用户界面(PUI)交互技术正是为为实现这一自然交互模式,而成为人机交互研究的焦点.通过对新兴的PUI领域及其重要分支——基于视觉界面(VBI)的介绍,阐述了利用头部定位与手的跟踪来实现VBI的方法,提出了在视觉方法中,要进行多方面的信息融合以实现更好的VBI.实验结果表明,基于视觉的感知用户界面(VBPUI)可以实现较高的

Graphics user Interfaces (GUIs) dominates the human-computer interaction (HCI) for some time, and has made that common people can use computer. However, in order to accommodate a wider range of applications, HCI needs to move toward interfaces that are natural, intuitive, adaptive, and unobtrusive. Perception User Interfaces (PUIs) is to implement this natural interaction and becomes a new focus in HCI. This paper first describes the emerging PUI field and the Vision Base Interface (VBI) which is a most important branch in PUI field. Then it expounds two methods which can realize VBI in real time:head locating which detect the position of head motion, hand tracking which fuse color information and motion information to control virtual object. At last it puts forward that not only visual method needs multi-channel information fusion so as to actualize good PUIs, but also PUIs needs many field (speech, haptic) fusion. Experiments show that Vision-Based PUI (VBPUI) can get better precision and real time utility and is a good method in HCI.