Emotion representation and physiology assignments in digital systems

Abstract

Emotions are of increasing interest to the HCI community. Within the last decade, emotion research in HCI grew from an eccentric hobby of some visionary scientists to a widely accepted field of research. A number of proof-of-concept prototypes and studies have been published, dedicated sensor systems and technology frameworks have been developed, and theoretical considerations have been made. While they all represent a very valuable contribution to this young field of research, they lack a common theoretical basis. Particularly, there exists no applicable model of emotions suitable for designing emotion-aware systems or performing HCI-related emotion studies. However, in order to become a mature discipline, emotion research in HCI needs such a rigorous footing that future work can be based on. In this paper, a suitable approach to structure and represent emotions for use in digital systems is introduced, after a detailed and critical review of widely used emotion models is given and representative study results are discussed. The proposed method meets several requirements of HCI researchers and software developers. It avoids artificial categorisation of emotions, requires no naming of emotional states, is language independent, and its implementation is straightforward. The results of an experiment based on this approach are discussed demonstrating its applicability

Emotions in context: examining pervasive affective sensing systems, applications, and analyses

“…The behavioral observations method is used to identify the user's emotional state by observing their externalized reactions, such as facial expressions and speech. Affect has numerous models of definition and typically follows two approaches: a categorical approach, which models affect as a distinct category, such as joy, anger, surprise, fear, or sadness, and a dimensional approach, which characterizes affect in terms of several continuous dimensions, such as arousal or valence [4].…”

Galvanic Skin Response: A Physiological Sensor System for Affective Computing

“…Because of this, the results are not as clear as for fear and anger. In paper [5], Erickson K. and Schulkin J. have found a decrease in heart rate, while paper [7] reported an increase of heart rate, while paper [8] could not find any significant difference to the measured baseline. Also, measuring positive emotions such as happiness seems to be very troublesome as can be observed especially in diastolic and systolic blood pressure and heart rate data [9], which can show a sign of decrease to even a strong increase.…”

Bio-sensing for Emotional Characterization without Word Labels

“…Peter and Herbon [6] have proposed a method for avoiding the use of words for automatic emotion recognition and provided guidelines about how to structure emotions as a dimensional representation for the use in human-machine interaction. Labeling emotions can be problematic because the category borders are blurry and the word 'anger' for instance can describe many different emotional states.…”

“…Labeling emotions can be problematic because the category borders are blurry and the word 'anger' for instance can describe many different emotional states. The method as described by [6] avoids these problems because it abandons labeling emotions with words. The procedure to classify emotions for automatic emotion recognition proposed in [6] consists of the following four steps:…”

“…Our current reported work is one of the first attempts to approach automatic emotion recognition with the novel method proposed by Peter and Herbon [6] that moves away from the notion of labeling emotional states with discrete categorical words. In the following section we describe which physiological modalities associated with emotions we chose to capture, the bio-sensors that we used, and the emotion elicitation method used with participants.…”