Facial Expression Recognition by Using Directional 2DPCA

In this paper a new technique Directional 2DPCA is used for facial recognition. Facial image firstly rotated in different angle, the directional 2DPCA that can extract features from the matrixes in any direction. In 2DPCA given the information in each row, but it cannot be uncovered the structural information. Facial features can be extracted applying any angel. Features were extracted from original facial image and it is rotated in six different angles. Then we sum a bank of Directional 2DPCA performed in different angle to develop multidirectional 2DPCA. The results of experiments on ORL and FERET database show that the proposed method is improved recognition rate than the traditional PCA and 2DPCA method.