SMU Guildhall Graduate Catalog Spring 2018 — Cohort 26 | Page 82

Wenzheng Huang Software Development « Using Computer Vision to Identify Player Inputs on Street Fighter 5 Machine Learning has become more and data from Street Fighter 5 with computer researchers, but researchers rely heavily going on in the game by simply “looking more popular among Artificial Intelligence on developer API (like StarCraft 2), memory hacking or reverse engineering to gain access to game status. Many at” it. It could be data-driven and apply to other scenarios with proper sample data. games are closed source, and do not I thought this project would provide an engineering or memory hacking the game data they want, and help them to better provide developer API, and reverse is strictly prohibited by end user license agreements. I spent about five months developing an application to identify and extract input 82 vision. This application can know what is SOFTWARE DEVELOPMENT alternative way for researchers to get the analyze behaviors of skilled players.