FERCE: Facial Expression Recognition for Combined Emotions Using FERCE Algorithm

Abstract

Research in facial expression recognition and emotion detection has undergone many developments recently. Human beings express different feelings via facial expressions at different circumstances. Perceiving such type of communication helps us in building a computational human cognition model. There are seven basic emotions: anger, disgust, fear, happy, sad, surprise and neutral. Researchers have stuck to those seven emotions for research into facial expression recognition and emotion detection. Apart from those basic emotions, the existing literature has defined 21 other compound emotions. However, they failed to explore some more emotions and few have been eliminated by claiming them as meaningless. This study proposes an algorithm that offers a new approach for grouping the emotion classes. In this proposed work a total of 37 emotions are identified as combined emotions out of which 16 are newly derived. These derived emotions were unnoticed by the earlier and they are existing in real-time human emotions. Along with exploring new emotions, the claim that certain existing emotions are symmetric is challenged and proved to be asymmetric in nature. This claim is validated by employing the Facial Action Coding System (FACS) and statistical analysis of combined emotions.