Evaluating the Effect of Emotion on Gender Recognition in Virtual Humans
Abstract
In this paper, we investigate the ability of humans to determine the gender of conversing characters, based on facial and body cues for emotion. We used a corpus of simultaneously captured facial and body motions from four male and four female actors. In our Gender Rating task, participants were asked to rate how male or female they considered the motions to be, under different emotional states. In our Emotion Recognition task, participants were asked to classify the emotions, in order to determine how accurately perceived those emotions were. We found that gender perception was affected by emotion, where certain emotions facilitated gender determination while others masked it. We also found that there was no correlation between how accurate an emotion was portrayed and how much gender information was present in that motion. Finally, we found that the model used to display the motion did not affect gender perception of motion but did alter emotion recognition.
Publication