复杂背景下的指尖检测方法研究电路与系统专业论文.docxVIP

AbstractAbstract Abstract Abstract Human and computer interaction has been traditionally restricted to the use of the keyboard，mouse or joysticks for a long time．Despite the variety of new devices， the friendliness and naturalness can not meet the users’commands．Vision．based gesture interaction becomes a hotspot in HCI field nowadays，which fuses technologies of image processing，machine vision，pattern recognition and other related disciplines．As a natural and intuitional input of man-machine interface，the gesture has wide prospect of applications in many fields，such as sign language recognition，game control and smart home etc．Fingertip detection is the key technology of many HCI systems based on the fingers’movement．However，some factors such as the dynamic diversity of backgrounds，unstable illumination and the random of finger movements，making finger’S precise positioning still a very challenging job． In this paper，we study techniques for fingertip detection under complex backgrounds．The main work Can be summarized as follows： 1、Three common object segmentation methods are studied in details，including skin color segmentation，background subtractions and image differencing． Experiments are carried out to analyze their characters and performance．Background subtraction，which focuses on updating the background model，is one of the most effective approaches when the background remains stationary．Image differencing Can adapt to the dynamic environment better，but it will fail when the object moves too slowly between two frames．It is proved that skin color is very important information which makes human distinguishable from others．However，it’S also not robust to illumination changes． 2、We study fingertip detection algorithms based on the shape area and curve analysis of the fingertip．The two approaches largely depend on the results of foreground segmentation，as both of them are operated after segmentation．The approach concerning the fingertip’S area character is sensitive to t