Adaptive Processing of Face Emotion Tree Structures

This paper describes a novel Recursive Neural Network for adaptive processing of FacE Emotion Tree Structures (FEETS). We proposed to use tree structures to represent Gabor Face Emotion Features. We demonstrated the robustness of our proposed system by testing against other well-known classifiers using the Cohn-Kanade AU-Coded Facial Expression database [1]. The system yields an accuracy of about 93% and 57% for known and unknown subjects respectively.