Research On Eye-Movement Based Intelligent Human-Computer Interaction Technology And Application

To meet the requirements of natural and intelligent interaction,new human-computer interaction technologies such as eye tracking,gesture recognition and EEG tracking have been developed continuously.People's vision contains rich implicit information,which can reflect the psychological state and visual preferences.Many scholars use eye tracking technology to mine and analyze potential personalized preferences from eye movement information.Eye tracking technology has the characteristics of direct,natural and sensitive,which can replace the traditional interactive input operation.Using eye movement information to control computers and intelligent devices can improve the naturalness and intelligence in the interaction process.Therefore,under the theoretical guidance of Control Science,Cognitive Science,Computer Science and Behavioral Science,eye movement data model and eye movement behavior model are constructed in this paper.Eye movement interaction mechanism and real-time recognition algorithms of eye movement behaviors are designed.Besides,eye control systems of virtual keyboard and unmanned platform are realized.Firstly,to reflect the temporal and spatial characteristics and trends of eye movement information,a spatial-temporal cube model(STC)based on eye movement data is constructed.Next,a region-based eye movement interaction mechanism is designed.Based on this,three eye movement behavior models are constructed:gaze,eye gesture and conscious blink.Real-time recognition algorithms of three eye movement behaviors are designed respectively.Then,according to the principles of eye movement interaction,a design framework of human-computer interaction system based on eye movement is proposed,and the interactive instructions are formally represented.Finally,the virtual keyboard eye control system and unmanned platform eye control system are built and eye control experiments are carried out.To avoid"Midas touch" and improve the performance of the interactive system,different eye control mechanisms are designed in terms of the characteristics of eye control systems.In virtual keyboard eye movement interaction experiment,the impact of different keyboard layouts on the interaction performance is analyzed,and the effectiveness of the proposed adaptive dwell time algorithm is verified.In eye-controlled unmanned platform experiment,the effects of two eye-gesture control methods on the interaction performance are compared and analyzed.