Facial expression classification

Nowadays, more and more advance electronic and machinery applications were invented to provide a better lifestyle to the society. Because of that reason, facial expression classification application also become important as it can help the electronic applications to interact with users in a more user-friendly method. Thus, a facial expression classification system using RBF neural network implementation is presented. As a beginning of the research in the facial expression classification, this project is done based on the shapes of the mouths. The mouths will be first undergone image preprocessing to obtain its shape and vectors. The vectors are needed for the neural network to process and learn to classify facial expressions. Radial Basis Function (RBF) neural network is used in this project as it provides advantages in pattern recognition. Networks are simulated for a few configurations and compared the result of testing. The results show that the percentages of correct matching are very high even though it is just based on the shape of the mouth. The percentage of correct matching can achieve in the range of 60% until 100%. Future improvements for facial expressions classification are suggested at the end of the project to improve the performance and functionality of facial expression classification in the future.