A supporting survey to step into a novel approach for providing automated emotion recognition service in mobile phones

Human face plays an important role to exhibit one's emotions. Beyond verbal communication, facial expression contributes much in human-to-human communication. In human-machine interaction system, emotion recognition plays a vital role. A wide range of techniques has been proposed to automate the emotion detection process in computerized systems. Basically six types of expressions are recognized, i.e., happiness, surprise, fear, sadness, anger, and disgust. The parts of face which reveal the emotion are eyes, eyebrows, lips, nostrils, muscles bulging area and wrinkles. The properties of images under consideration are grayscale or colour image, strong or light background and the lightings. The source image can be a photo captured from the camera or it can be the image from video stream. There are three stages of processes involved in this emotion detection system which are face detection from the given image, extracting its features, and classification. The techniques involved in these three major processes and their sub processes are analyzed in this survey paper. Some of the soft computing techniques are also discussed to train the system to fine tune the system. The datasets utilized in various techniques includes JAFFE dataset, eNTERFACE database, Cohn Kanade AU-Coded database, MMI facial expression database, Berlin emotion database, CK+ database, and Chen generated dataset. This paper aims at identifying the techniques for recognizing facial emotions which are suitable for mobile phones.