Hierarchical multi-stream recognition of facial expressions

Achieving optimal recognition accuracy, particularly under conditions of input data variability, is a challenge for automatic facial expression recognition. However, little research has been devoted to investigating the robustness of automatic expression recognition under adverse conditions. This paper investigates an approach for recognition of facial expressions that is founded on hierarchical state-based modelling of streams representing spatially localised expression dynamics. Experimental assessment shows that the proposed model achieves high and stable recognition accuracy over a range of input data degradation. Moreover, inter-stream coupling is shown to make a significant contribution to recognition accuracy.