Automatic detection of emotion from real-life data

Recognition of emotion in speech has recently matured to one of the key disciplines in speech analysis serving next generation human-machine communication. This paper provides the best practices in the automatic detection of real-life emotion from vocal expression. “Real-life” emotion is hard to collect, ambiguous to annotate, and tricky to distribute due to privacy preservation. Acting of emotions was often seen as a solution to the desperate need for data. In contrast with most previous studies, conducted on artificial data with archetypal emotions, this paper addresses some of the challenges faced when studying real-life non-basic emotions. What needs to be done in this field to improve emotion detection is also discussed.