Story Generation System Using Player’s Emotions for Review in Gamed-Based Learning

Abstract

We propose a game-based story generation system that automatically generates scripts in real time by using a player’s emotions and actions. In this paper, we consider whether the history of the experience, containing action and emotion information, would be useful for review of the learning experience. The system provides a player with a virtual world and a virtual tool operated by using a hand controller. The system recognizes the player’s real-time emotions through facial expressions, and it outputs reactions based on these emotions via knowledge-based systems when the player operates the tool. Then, it outputs scripts based on the emotions and the history of actions. We evaluated the system by conducting experiments with university students as subjects. As a result, the system could generate interesting stories based on the player’s experience in the game by using the player’s history of actions and emotions. The results suggest that the historical information, including the learner's real-time actions and emotions, is helpful for review in learning.