For face recognition, it is very important determining which features of the faces will be used in the classification process. The identification based on appearance uses the pixels of the corresponding image to extract the features. Using the pixels directly is not very efficient due to the high dimensionality of the resulting features which results in a poor discriminative capability between different persons and in increased computational complexity. Implementing any kind of data transform could be a good strategy for reducing the dimensionality of the data and increasing the discriminator capability. Using PCA or DCT transforms it is possible to implement systems with a good rate of recognition if the number of recognizable persons is low. In this project, it has been investigated others features extraction techniques, especially the ones based on Local Binary Patterns.