Computer Vision

I thought I'd share some of my school work with you. As I previously mentioned, I'm in a computer vision class that has been really interesting. We've had 4 assignments this semester and each one has taught us more about certain areas within Computer Vision. Our last assignment was to create a face tracker using a techniques we learned in class called particle filtering. So, after I completed the assignment and turned it in, I thought it would be fun to let other people see what I am doing, so I exported the images to a video and then uploaded it so you could see it. Here it is, I will explain a little more after the video.


Reid's face tracker

The orignal video was provided to us by our professor. The man in the video is a professor at McGill University in Montreal. The dots on the video represent the "particles" that help track the face. Each dot gets tested to see how likely it is the face. This "likelihood" is represented by the brightness of each dot on the video. I then use the average of these points to estimate where the face probably is. That is represented by the red dot and rectangle.

As you can tell it does an OK job of tracking the face over the course of the video. It isn't perfect but I'm pretty proud that it works at all.

Anyway, let me know what you think.