Facial images based emotion classification

U radu je implementirana metoda za prepoznavanje emocija na slikama lica. Korišteni su Gaborovi filteri kako bi se dobile crte na licu i potom PCA kako bi se smanjila dimenzionalnost. Kao klasifikator koristi se SVM. U teorijskom dijelu objašnjeni su pojmovi kao što su emocija i afekt. Opisan je klasifikator i korištene metode. Praktični dio se sastoji od programa za dohvaćanje značajki iz baze slike i od dijela koji koristi te značajke kako bi se istrenirao klasifikator te kasnije koristio prilikom prepoznavanja emocija, što s korisnikove kamere, što s učitane slike.A method for recognizing emotions in images of faces is implemented in the thesis. Gabor filters are used in order to extract facial features and then PCA is applied afterwards in order to reduce dimensionality. SVM is used as a classifier. In the theoretical part, terms such as emotion and affect are explained. The utilized classifier is described, as well as the applied methods. The practical part consists of an application which extracts features from a base of images and from a part which is using those features to train the classifier in order to use it later on when recognizing emotions in an image from a personal camera or from an uploaded one