Differential components of discriminative 2D Gaussian–Hermite moments for recognition of facial expressions

摘要: Abstract This paper deals with a new expression recognition method by representing facial images in terms of higher-order two-dimensional orthogonal Gaussian–Hermite moments (GHMs) and their geometric invariants. Only the having high discrimination power are selected as set features for expressions. To obtain differentially expressive components moments, discriminative GHMs projected on to expression-invariant subspace using correlations among neutral faces. Features obtained from used recognize an well-known support vector machine classifier. Experimental results presented commonly-referred databases such CK-AUC, FRGC, MMI that have posed or spontaneous expressions well GENKI database has in-the-wild. Experiments mutually exclusive subjects reveal performance proposed is significantly better than existing similar methods, which use local patch-based dimensional binary patterns, directional number patterns generated derivatives Gaussian, Gabor- other moment-based features.