Sparse representations for facial expressions recognition via l1 optimization

In this paper, the principles of sparse signal representation theory are explored in order to perform facial expressions recognition from frontal views. Motivated by the success such methods have demonstrated in the face recognition problem, we formulate the feature extraction procedure in order to achieve facial expression recognition as an l₁ optimization problem. We show that the straightforward application of these methods to expressive images imposes certain difficulties. The use of difference images (i.e., the images that are derived from the subtraction of the neutral image from the expressive one) for sparse facial expression representations is justified. The use of expressive facial grids for similar tasks is also studied. Finally, the robustness of the proposed representations under facial image occlusion is shown and the efficacy of the proposed method in a series of experiments is demonstrated.