Face Detection by Using OpenCV’s Viola-Jones Algorithm based on coding eyes

Abstract

Facial identification is one of the biometrical approaches implemented for identifying any facial image with the use of the basic properties of that face. In this paper we proposes a new improved approach for face detection based on coding eyes by using Open CV's Viola-Jones algorithm which removes the falsely detected faces depending on coding eyes. The Haar training module in Open CV is an implementation of the Viola-Jones framework, the training algorithm takes as input a training group of positive and negative images, and generates strong features in the format of an XML file which is capable of subsequently being utilized for detecting the wanted face and eyes in images, the integral image is used to speed up Haar-like features calculation for each image in (MIT, FERET) dataset and the adaboost algorithm is implemented to collect the weak classifiers and produce strong classifier. By using classifier cascade process, the speed and accuracy of face detection system is increased .The proposed method has accuracy is about 98.97% for detection faces.