Cultural dimension in emotion recognition for human machine interaction.

Maria Alejandra Quiros-Ramirez, Takehisa Onisawa

Abstract: Human emotion recognition is a multidimensional task. In this paper we study the effect of the cultural dimension in emotion recognition models and their use in human computer interaction. We prepared two experiments to analyze the consequences of disregarding culture in emotion recognition models. The results show that failing to consider the user's culture while applying emotion recognition techniques in interaction scenarios decreases the system's performance, making the emotional input meaningless and detrimental to the system and interaction.