Gesture Recognition And Human-Computer Interaction System Based On Skin-Color Segmentation And Statistical Template Matching

In recent years,with the rapid development of computer technology,more natural and efficient new forms of human-computer interaction continue to emerge.Gesture is one of the basic ways of human communication,which conforms to the daily communication habits of mankind.Based on monocular vision technology,it is a hot area of research to realize human-computer interaction by means of gesture recognition.At present,although lots of algorithms of gesture skin-color segmentation and gesture recognition are proposed,the existing algorithms should be perfected because they have shortcomings in the recognition rate,executing efficiency,practical applicability and so on.For example,most of the static gesture recognition algorithms are highly complex,and it is difficult to achieve the ideal gesture segmentation effect in environments with complex background or poor lighting conditions,resulting in low gesture recognition rates.In view of these problems,this paper focuses on theoretical and applied research on gesture color segmentation and static gesture recognition.The main work and contributions are as follows:1.A multi-factor gesture skin color segmentation method is proposed on the basis of analyzing the related technology.By using this method,the skin color is firstly segmented by elliptical skin color model,and then the background model is established by using the moving object detection method to eliminate the area of skin color in the background.Next,the face color recognition region is excluded from the face color recognition region.The experimental results show that the proposed method can obtain better gesture segmentation in the complex background or poor lighting conditions.2.A simple and effective statistical template matching algorithm is proposed to realize static gesture recognition.Firstly,based on the normal distribution probability model,the statistical template features corresponding to various gestures are generated by using the collected gesture image samples.Secondly,the similarity between gesture images is defined by template feature,and then the matching judgment rules are used to distinguish gestures to determine the gesture category corresponding to the gesture image to be recognized.The experimental results of 11 kinds of gestures show that the algorithm proposed in this paper can obtain the average recognition rate higher than 93.5%,which is superior to the existing similar algorithm.3.In this paper,the gesture recognition and gesture recognition algorithm are applied to human-computer interaction,and a gesture recognition interactive system is designed and implemented in C ++ as programming language,combined with MFC development framework and OpenCV.The system provides 11 kinds of gestures,which can be applied to simulate the mouse and keyboard operation,so as to control the operation of the PPT,the player and other software.This system provides a friendly interface,and with this system the implementation efficiency is high,so the system is of high versatility.