Cheer Me!: A Video Game System Using Live Streaming Text Messages

Abstract

Nowadays, video game live streaming has gained popularity. To make communication between the streamers and the viewers more direct, we developed a game system in which the viewer can participate and reflect their own emotion by posting chat messages on a live streaming service. The system evaluates and classifies the emotion of the viewers’ messages into nine levels through a text analysis API. The smaller the level, the more negative emotions the sentence has, and the higher the level, the more positive they are. When the system evaluates the emotion of the message, the message sentence appears on the game screen in real time. The sentences have different effects, which help or bother the streamer’s play depending on classification. In a user test, five subjects were interviewed. All of them answered that they felt a strong awareness of participation and that they thought that the system activated communication between the streamer and the viewers. A “big hole” trick, which was designed so that player could pass the scene only when he could receive the viewers’ messages, gave them excitement.