Download now Free registration required
Recent work in mouse movement analysis has determined that, with sufficient data, users can be uniquely identified solely by their mouse movements. This paper considers the domain of video games and attempt to use mouse movements to identify game players. The authors conduct a user study that requires users to perform baseline tasks in a controlled environment, and then play Solitaire and StarCraft, two popular video games. This paper extracts features from the mouse movement raw data and employs basic (k-Nearest Neighbor) and complex machine learning techniques to classify users. The accuracy of the identification varies with the variety of moves available in a particular game.
- Format: PDF
- Size: 516.1 KB