Face detection and recognition methods

Abstract

Foto analizinin en onemli uygulamalar ndan biriside cesitli metodlar kullan larakyap lan yuz tan ma islemidir. Özyüz tan ma yöntemide bu metodlardan birisidir ve resimleraras ndaki farklar tan mlamak icin o resimlerin küçük gruplar halindeki karakteristikresimlerini kullan r. Buradaki amaç da l m n kovaryans matrisinin özvektörlerini yaniözyüzleri bulmakt r. Tipik resim boyutlar için bu özyüzleri bulmak oldukça zordur ve buyüzden pratik uygulamalarda yak nsamalar yap labilir. Tan ma islemi, yeni resmi özyüzlerinolu#turdu u uzaya projeksiyonu ile yap l r. Bu projeksiyon sonucunda, yeni resmin pozisyonitibar yla en yak n oldu u yüz seçilir.Özyüz tan ma yöntemi kolayl , h z ve kolay ö retilebilirli i yüzünden yeterli ve iyibir yöntemdir.Bu projede özyüz yöntemi kullan larak bir yüz tan ma i#lemi gerçeklenmi#tir.

One of the most succesful applications of image analysis and understanding is facerecognition that can be obtained by using several methods. Eigenface method is one of thesemehods in which a small set of characteristic pictures are used to describe the variationbetween face images. Goal is to find out the eigenvectors (eigenfaces) of the covariancematrix of the distribution, spanned by a training set of face images. Later, every face image isrepresented by a linear combination of these eigenvectors. Evaluation of these eigenvectorsare quite difficult for typical image sizes but, an approximation can be made that is suitablefor practical purposes. Recognition is performed by projecting a new image into the subspacespanned by the eigenfaces and then classifying the face by comparing its position in facespace with the positions of known individuals.Eigenfaces approach seems to be an adequate method to be used in face recognition dueto its simplicity, speed and learning capability. In this project a face recognition system, basedon the eigenfaces approach is proposed.