Novel higher-order local autocorrelation-like feature extraction methodology for facial expression recognition

A novel feature extraction method for facial expression recognition from sequences of image frames is described and tested. The authors propose HLAC-like features (HLACLF) for feature extraction. The features are extracted using different masks from Grey-scale images for characterising facial texture. Then the most informative features are selected based on mutual information quotient (MIQ) criterion. Multiple linear discriminant analysis (LDA) classifier is adopted. The proposed system is fully automatic and including: face detection, facial detection, feature extraction, feature selection and classification. Experiments on the Cohn-Kanade database illustrate that the HLACLF is efficient for facial expression recognition compared with other feature extraction methods.