A Bimodal System for Emotion Recognition via Computer of Known or Unknown Persons in Normal or Fatigue Situations

The recognition of emotion/mood in groups of people is particularly interesting as it constitutes a whole new research area that can provide solutions to various high level problems and applications, such as, for example, the development of educational technologies for groups of learners in which teaching is adjusted to the mood of the group. The purpose of this chapter is to summarize and present research on the recognition via computer of emotions and/or the mood of a single person or groups of people by using a visual-facial modality combined with a keyboard-based modality. Additionally, this chapter presents research about a program that recognizes the emotions/mood of groups of people in live pictures collected directly from webcams. New algorithmic approaches are proposed and conclusions are drawn for use in similar systems. The implemented system successfully recognizes emotions in the following situations: 1. The system is only trained to recognize emotions/the mood of a specific person. 2. The system is trained with one set of sample emotions/mood from which it aims at generalizing and recognizing emotions/mood of unknown people. 3. The system is trained with several sets of sample emotions/mood from which it aims at generalizing and recognizing emotions/the mood of unknown people. The chapter is complemented with illustrations of use of the implemented system, as well as suggestions for future related work in this area.