Image Based Event Forecast

OData support
Dr. Vajda Ferenc
Department of Control Engineering and Information Technology

Nowadays, computer vision and image processing is one of the fastest growing areas of electrical engineering and computer science. Due to the appearance and spread of the digital video recording equipment and developing of image processing the tasks which were complicated and insolvable before have become routines. The complicated and expensive devices used earlier can be substituted by camera-equipped mobile devices.

Gambling has been played for thousands of years and it is still very popular today. One of the most famous casino games is the Wheel of Fortune. In this game you have to guess a sector of a rolling wheel and if the wheel stops there you win your bet multiplied but if you don’t make a lucky hit you lose your money.

Although the outcome of the game is uniformly distributed ideally, that is every sector has the same chance for winning, the wheel’s movement can be described precisely with physical equations. Consequently, if the wheel is moving and the initial values (speed and position) are known, the game’s outcome can be predicted. In most versions of the game, player is allowed to bet after the wheel has been started.

In my dissertation, I have designed and built up a computer vision system, which can predict the final position of the wheel, after a few turns. Results of testing proved that the prediction is correct in over 40 % of games, which obtains an expected return of over 500%, in contrast, the expected return for an uninformed gambler is −6.25%.


Please sign in to download the files of this thesis.