Facial Expression Recognition Based on Combination of Spatio-temporal and Spectral Features in Local Facial Regions

This paper presents two new approaches for facialexpression recognition based on digital curvelet transform andlocal binary patterns from three orthogonal planes (LBP-TOP)for both still image and image sequences. The features areextracted by using the digital curvelet transform on facial regionsin still image. In this approach, some sub-bands correspond toangle of facial region is used. These sub-bands consist of morefrequency information. The digital curvelet coefficients and LBPTOPare represented to combine spatio-temporal and spectralfeatures for image sequences. The obtained results by ourproposed approaches on the Cohn-Kanade facial expressiondatabase have acceptable recognition rates of 91.90% and88.38% for still image and image sequences, respectively.

Keywords:

facial expression recognition , digital curvelet transform (DCUT) , local binary patterns from three orthogonal planes (LBP_TOP)