A perceptual evaluation of social interaction with emotes and real-time facial motion capture

Social interaction between players is an important feature in online games, where text- and voice chat is a standard way to communicate. To express emotions players can type emotes which are text-based commands to play animations from the player avatar. This paper presents a perceptual evaluation which investigates if instead expressing emotions with the face, in real-time with a web camera, is perceived more realistic and preferred in comparison to typing emote-based text commands. A user study with 24 participants was conducted where the two methods to express emotions described above were evaluated. For both methods the participants ranked the realism of facial expressions, which were based on the theory of seven universal emotions stated by American psychologist Paul Ekman: happiness, anger, fear, sadness, disgust, surprise and contempt. The participants also ranked their perceived efficiency of performing the two methods and selected the method they preferred. A significant difference was shown when analyzing the results of ranked facial expression realism. Happiness was perceived as the most realistic in both methods, while disgust and sadness were poorly rated when performed with the face. One conclusion of the perceptual evaluation was that the realism and preference between the methods showed no significant differences. However, participants had higher performance in typing with emotes. Real-time facial capture technology also needs improvements to obtain better recognition and tracking of facial features in the human face.