Design And Implementation System Of Human Computer Interaction Based On Single Camera

Computer vision is a challenging research subject in the field of pattern recognition recentyears, specifically, the image sensor sensing image processing to make a meaningful judgmenton the actual objects and scenes, with the popularity and continuous development of computervision technology in electronic products, the growing number of scholars and experts arecommitted to the human eye detection technology research in the field of computer vision. Eyedetection technology and other related technologies combine will better serve the people’s lives,the human eye detection technology and machine vision technology to combine the driverfatigue detection system. Human-computer interaction systems based on eye trackingtechnology. And virtual reality technologies combined with virtual gaming interface.This study is divided into face detection, eye pupil detection, corner of the eye pointdetection, human-computer interaction system design four main parts. Human eye detectionalgorithm on the basis of the origin analysis, and the system of human-computer interactionsystem algorithm is verified by experiment for a single camera, established a line of sight as thesystem input signal of human-computer interaction system and VC++6.0developmentplatform.For the face detection, location algorithm for eye pupil based on skin color features andGenetic Algorithm was presented, in the YIQ and YUV space of human skin color pixels for alarge number of statistical analyses, the human skin color in the space show better clusteringcharacteristics, the characteristics of the face region segmentation as a criterion, at the sametime, the color space of the lip color characteristics as the face region segmentation is anothercriterion, according to the two criterion, combined with the region growing algorithm for facedetection. Solve the problem of face detection in the case of light and head deflection.The basis of face detection, using the face region positioned, according to priorknowledge the search area was narrowed, in the face of the upper left the subregion usinggenetic optimization algorithm of eye area detection, for the human eye pupil center positionis to use the geometry of the round the property positioning, for the detection of corner pointsthis paper, harris corner detection algorithm in the corner response function, with varianceprojection function to locate the corner points, in human-computer interaction, the center ofthe pupil as a fixed point, corner of the eye point as a moving point, will move the distancebetween the moving point and fixed point map is transformed to the choice of the sub regionon the interactive interface, realize human-machine interaction interface design, experimentalresults indicate as follows: positioning the eye has a good performance in the complex background and head deflection. Solve the problems of the applicability of the traditionalalgorithm for the detection of different human eye.The development human-computer interaction system based on single camera, real-timedetection of the human eye, by the length of the human eye fixation time on the computerscreen as the instruction of the human-computer interaction to human-computer interaction,the human-machine interactive system hardware does not too complex, the system has betterrobustness and real-time in the head rotation and complex background, the system iscompared with the current mainstream human-computer interaction systems, low cost, easy touse, easy to universal characteristics.