My Acting Pedigree

For a course project, a friend of mine built a Wii Gesture Classification system. The idea is that during game play, a user will perform actions with the Wiimote which have to be identified by the game and translated into actions by a game character. These actions have to be classified properly. For example, consider a baseball game in which a full swing needs to be distinguished from a bunt.

The difficulty with building such a system is that fact that there will be a significant amount of variation in the way different people will try to perform the same action. For example, people swing a baseball bat in many differenet ways, and even do so from different sides. Moreover, as different games will require different action sets to classify, it would be nice to have a system that could be re-used in multiple games without much effort.

My friend's system is trained through the use of examples. To do so, a number of people perform the desired actions. The system then attempts to generalize from these training examples to a more general idea of the actions. The resulting model is then used to classify actions by a game player.

This is all a long-winded way of getting to the point that it is a cool system, he made a demo movie of his software for the 2009 AAAI Video Competition, and I am in it. In the video, my friend shows the system classifying actions in real-time. Since I can't resist the allure of movie-making, I volunteered to be one of the test subjects.

Anyways, here is the video:


You can also find contact info if you are interested in the system at the following link: http://wiigesture.com/

An Update

Since I am currently writing my thesis during the day, I find I am less inclined to come home and write on this here blog. Hopefully I will be able to satisfy my legions of fans until July when I will be more interested in writing. Until the next time I feel obligated to write anything, I leave you with the following updates/observations.

I will begin by saying that I have decided to grow a thesis beard. Specifically, I will not shave until my thesis writing is complete. There are a number of reasons for this decision. First, thesis writing is like the playoffs of a graduate degree. Secondly, looking in the mirror will be a constant reminder of how lazy I am. If I had finished my thesis sooner, I wouldn't look this bad. This should provide extra motivation for finishing as soon as possible.

Finally, the thesis beard will allow me to officially be better than Sidney Crosby at something. He has millions of dollars, is a world-class athlete, and is well-spoken. I am none of these things. But at least for the next few years, I can still grow a better beard than him. Sure, I assume I am better at mathematics and computer science than he is, but I don't know for sure. He may just be hiding those skills. However, concerning beards, I have him beat. At least for now.

A second update: I have run out of toothpicks. When I arrived in Edmonton 20 months ago, I purchased 500. Today, I ran out. Considering that I did not spend that entire time in Edmonton, I am guessing I am going at about a toothpick per day pace. Make of that what you will.