基于人眼检测优化的立体人机交互技术的研究

3D human-machine interaction based on human eye detection

随着3D显示技术的逐渐成熟,越来越多基于3D显示技术的产品被应用于各个领域,例如3D电影、裸眼3D电视、体感交互式系统和影像教学等。基于3D显示的人机交互系统也是近年来越来越热门的研究方向,但是却没有高效地利用深度信息作为交互的线索,因此本文提出将3D显示与体感交互结合,实现立体人机交互的虚拟键盘,让用户只需点击屏幕前方键盘的图像即可进行对应按键信息的录入和编辑,其帧率可达10.6fps。针对3D显示中用户视角偏离会导致按键的像在视觉上的偏移,容易造成误输入的问题,在系统中整合人眼检测功能,修正因视角偏离带来按键定位误差,从而实现了准确率100%的字符输入交互功能。

With the growing of 3D display,more and more products based on 3D display are used in various fields,for example,3D movies,auto-stereoscopic television,3D surgery,etc.Human-computer interaction(HCI)system based on 3D display is also a popular research direction in recent years.However,most of these systems don’t use depth information as a clue for interaction which limit interaction orientation into a 2D plane.To extend control on depth direction,a novel HCI system based on 3D display is presented.The system shows a keyboard style interaction interface by a 3D monitor,and uses binocular cameras to acquire the gesture and positions of the user’s hand and fingers which can input information to the computer.And the system input frame rate can reach 10.6 fps.For problem that the movement of the user will cause the image of the button in his vision to be shifted correspondingly,which may cause incorrect input the human eye detection is integrated into the system.By the optimization of user’s viewing angle calibration,we can correct button’s positioning error,thus achieve a great accuracy of character input 100%.