Children Emotion Recognition Approaches, Databases and Applications

Abstract

With the advances in the field of emotion recognition, increasing number of investigations have been done and they have found a significant place in different applications involving human computer communication. The primary objective of an emotion recognition system is to interpret the input signals from different modalities and use them to convey information about the interpreted emotion. For example, the doctor may act based on patient's emotions recognized. Autism Spectrum Disorder children’s primary emotion is hard to be recognized and identified from their speech. Being a multidisciplinary field involving computer science, psychology and cognitive science, AC has instigated the computer scientists in building up emotion detecting models to promote human-machine interfaces. A typical Autism is a lifelong disorder and it is difficult to find out an emotion from their speech through signal processing. People can recognize happy and sad emotions from such children’s speech between the ages of 5-10 years. Emotions in humans are expressed in different modes which gives a lead to the development of affect recognition systems. Different modes, individually or in various combinations contribute to diverse approaches for emotion recognition. In this paper a novel method is presented for ASD children’s emotions from their speech.