Face Detection Based On OpenCV Vision Library

Face detection computer technology which can find the position and figure ofpeople’s face. It detects facial features and ignores interference factors includingconstructions, tress and clothes. For some cases, face detection can also figure out thesubtle features on faces, such as the exact poison of eye, nose and mouth. Due to thegreat contributions face detection has made in the field of security testing system,medicine, files management, video conference and man-machine interaction, itsapplication future becomes a research hotspot of current pattern recognition andartificial intelligence.This paper designs and realizes the detection of faces on thebasis of OpenCV vision. Using the binding model of arithmetic Adaboost, it achievesface detection by Haar feature selection. It plays an extremely important role in thewhole software because image processing has a direct impact on the accuracy rate ofpositioning and identifying. The major approach of image processing of this softwareis the image representation technique of summed are table and a cascaded taxonomicstructure on the basis of Haar and the principle and training process of arithmeticAdaboost. In the process if identifying, firstly train the haarcascade_frontalface_alt2of cascade classifier OpenCV though arithmetic Adaboost and then detect andposition the sample faces by classifier. After the codes are designed and debugged,face detection and positioning finishes its task perfectly when testing digital images inthe final examination, improving the accuracy of positioning and identifying.