Creating face recognition system using template matching and multilayer perceptron

In this thesis, face recognition systems has been carried out using different methods. In order to test the methods, CNNL database has been used. CNNL database includes 10 different expressions of each person that belongs to 1266 people. CNNL database has been composed of 2 sections which are memory set and test set. Test set includes 2 different expressions of each person. Memory set includes 8 different expressions of each person.Firstly, template matching method has been used. Face recognition system with template matching method has been carried out using City Block Distance, one of the Minokowski class metrics.Secondly, in order to boost the value of success rate which has been obtained in the previous section, 2 different methods has been used. In the first method; second template matching method has been applied by extracting some parts of the image which decreases the success rate. In the second method; pupils has been automatically detected and the distance between 2 pupils has been measured. Success rate has been boosted with this distance.Thirdly, success rate has been examined using Multi Layer Neural Networks.In the last section, image dimensions has been reduced and the performance of the system (success rate and process time) has been examined.