Currently, the football analytics staff gathers the video and manually tracks data points from the video. It is a tedious process that takes too much time. With only 6 days to work between games, it is very difficult to finish recording data in time to use for game preparation.
What it does
We sought to create a program that could import a video and use machine learning and computer vision programs to determine data points automatically for the football team's data analysts.
How we built it
We create a python script to import the video and scrape the multiple images. Through those images, we classify if the image is game film in side view or field view. We then used these classifications to update an excel file with info that the football team's data analysts would normally have to upload manually.
Challenges we ran into
Time was definitely an issue while working on the machine learning aspects of our project. Learning to use different packages was also a challenge. The most difficult thing was putting all of our ideas/skillsets together to make a working product.
Accomplishments that we're proud of
Working with our different backgrounds to solve a big problem to hopefully help Georgia Tech be successful.
What we learned
Solving problems in 24 hours is tough, but we can conceptualize the idea to solve the problem and create a framework that hopefully allow
What's next for Bobby Dodd Squad
We hope to get a call from Coach Collins to join forces to help beat U(sic)GA next year. In reality, we would like to advance this program to identify other data points we did not have time to program at this time. We would look at different machine learning and computer vision technologies out there to see what other technologies are out there to help identify other data points that could be collected automatically.