Fuzzy Logic for Video Game Engagement Analysis using Facial Emotion Recognition

Computer games have become popular recently, yet evaluating the user’s emotional state to understand their engagement lags far behind. There are few methods of assessing emotional responses and even fewer methods of quantifying emotions during play. This paper proposes a method for classifying a player’s emotional state and quantifying the engagement of a player in a game. The Emotions are recognized using a facial emotion recognizer built using convolutional neural networks. A fuzzy logic model transformed the seven universal emotions into low, medium or high engagement. This approach offers a method for quantifying emotional responses during a play experience.