Survey of mood detection through various input modes

Abstract

Mood has a large impact on people's behavior and even health. Thus, detecting and monitoring mood can potentially benefit users, researchers, clinicians, and content providers. In recent years, advancements in affective computing have enabled the development of various mood detection systems based on self-reported data, speech, facial expressions, mobile phone usage patterns, or physiological signals. This paper reviews each of those approaches and evaluates them in terms of usability and accuracy. Systems based on mobile phone usage and physiological data seem to be the most user friendly, but more research is needed to examine the positive and negative effects of mood monitoring.