The analysis of emotional expressi̇on development of face transplant patients by using image processing algorithms

In this thesis, appearance based Gabor Wavelet Transform and Local Binary Pattern methods were adopted for recognition neutral and 6 emotional facial expressions consisting of angry, scared, happy, hate, confused and sad. The healthy group and two face transplant patients were joined. Using both methods and serially combination of this methods feature extracted. Adopted approaches were performed firstly on highdefinition database which we form. We focused eyes and mouth location which are the most prominent regions while performing expressions. Eyes and mouth region features were subjected to evaluation individually. Afterwards, the dimension of images in the database was reduced. The same methods were applied in these images. Also feature vectors were composed from the combination of eyes and mouth features and were given to K-Nearest Neighbour input. At first step the healthy group images were used for training and testing, at second step images of face transplant patients were used for the test input of trained system while classifying. Namely regional facial mimics of face transplant patient were examined by comparison the healthy group. Finally the results obtained for the normal group and transplant patients were evaluated. Apart from all, results obtained by the proposed method were classified by comparing among themselves and results were evaluated.