A Hybrid System Based on Wrinkles Shapes and Biometric Distances for Emotion Recognition

Communication has an important role in human interaction. It can be vocal, bodily, textual or emotional in order to help everyone to understand others and be understood. Generally, all these types of communication aim at expressing a thought or an idea via a set of selected words, gestures, sounds or emotions. Emotions are so important to express ourselves and understand the others even without words or sounds. This paper presents a hybrid emotion recognition system based on wavelet network using 1D fast wavelet transform. The proposed application is based on two approaches. The first one is based on the shapes of the wrinkles; the second is based on the biometric distances. We combine these two approaches in order to ameliorate the classification rates. The rates given by experimental results show the effectiveness of our proposed system.