Augmented Reality Game

This was a group project created for one of my classes while attending UOIT. We took the opportunity to look into the world of augmented reality and decided to create a card game based prototype for the existing game Pokemon. The group members consisted of Vincent Bui, Andrew Milner, Richard Pitul as well as myself.

Screen Sample of Current Prototype

The prototype has full 3D model support including animations (.FBX), voice recognition for calling out attacks (Microsoft Speech SDK), emulated 3D surround sound support (FMOD) and camera pattern recognition (ARToolkit). On the non technical side basic gameplay works; one can play the game until one of the Pokemon are "knocked out". This means turn determination and damage calculations are all functional.

I was responsible for the integration of the FBX model parser, damage calculation, gameplay loop and a system that determined which and when different animation loops needed to be played. This project was fully coded in C++. A more detailed explaination of how the project was made can be found in the presentation paper at the bottom of this page.


There were a couple of issues that came up during the development of this project. One particular issue was the resolution of the webcamera we were using for the pattern recognition. If the camera is not sharp enough to detect the patterns then there will be an issue with the placement of the models on the markers. This issue limits the distance the camera can be placed away from the cards limiting the actual play area.


Another issue with the pattern recognition software is that if at anytime that the black box is blocked by any object then the pattern will not be recognized by the software. This means that the patterns need to be as flat as possible with no creases or blemishes. Glare from the surrounding environment can also be an issue with the pattern recognition. The camera position also is an issue. The camera needs to be set up at an angle where the pattern is not too narrow to be detected. This will either result in a non-recognizable pattern, or it may read it as a different pattern entirely, resulting in the wrong graphical output.

The research paper can be downloaded provided by the link below:

All content is copyright © 2015 Sing Chen, unless otherwise specified. All Rights Reserved.