Abstract Expressions of Affect

摘要

What form should happiness take? And how is disgust shaped? This research investigates how synthetic affective expressions can be designed with minimal reference to the human body. The authors propose that the recognition and attribution of affect expression can be triggered by appropriately presenting the bare essentials used in the mental processes that mediate the recognition and attribution of affect. The novelty of the proposed approach lies in the fact that it is based on mental processes involved in the recognition of affect, independent of the configuration of the human body and face. The approach is grounded in (a) research on the role of abstraction in perception, (b) the elementary processes and features relevant to visual emotion recognition and emotion attribution, and (c) how such features can be used (and combined) to generate a synthetic emotion expression. To further develop the argument for this approach they present a pilot study that shows the feasibility of combining affective features independently of the human configuration by using abstraction to create consistent emotional attributions. Finally, the authors discuss the potential implications of their approach for the design of affective robots. The developed design approach promises a maximization of freedom to integrate intuitively understandable affective expressions with other morphological design factors a technology may require, providing synthetic affective expressions that suit the inherently artificial and applied nature of affective technology.