Birleşik yerel ikili örüntüler ve yerel faz kuantalama kullanılarak yüz ifadelerinin tanınması

Face expressions have a significant role in the interactions among the people and they are primary means of conveying the social information. Automatic perception of face expression by computer software provides solution for the necessities in several application areas. The primer areas where there is a need for face expression recognition are:safety, health, education etc. Generally, face expression recognition systems consist of face detection, extracting of features, and classifying emotion steps. Establishing a powerfull system of recognition of facial expression producing reliable results, distinguishing characteristics of face must be extracted from related parts of face. Therefore, feature extraction method has a great importance. In this thesis, Local Binary Patterns (LBP) and Local Phase Quantization (LPQ) methods which provide prominent facial features extracting are used together as feature extraction methods. Also, it is aimed to increase the classification success rate by benefiting from the complementary features of these two methods. With this thesis, it is aimed to find an answer to the question that "Does combining LBP and LPQ contribute to success rate of classification of face expressions?". As result of experiments conducted, the success rate of classification has contributed by using combined LBP and LPQ methods. This success rate is found 88%.