Day 8

The day began with the usual meeting, where I learned a little about what the other lab groups are doing and what their final objectives are. After the morning meeting, we split up into our lab groups  and I began to code the program that would work through the perspective transformations. I finally completed it right before lunch after working on it for almost 3 hours. After running some numbers through it, most of my x and y values were within 2 feet of the known values for those points. Joe came in and ran through a list of things that needed to be completed and a relative time to have them completed by so that we are on track to accomplish the final objective. We should be able to come to a conclusion on the accuracy of mapping players in the broadcast footage to the hockey rink by the end of the week, and then should spend the next two weeks collecting data on the accuracy of identifying which team the player is on and associating those players with names and numbers. After that, we should do some brainstorming to answer some other less rigorous questions and start working on the final presentation. After going over this outline, Joe found the video of the footage from the center fixed focal length camera that was paired with a video of the players location on the rink. Anoop and I began to use our program to determine our program's accuracy versus the accuracy of the freshman method. Most of the x-values seemed to be spot on, but the y-values seemed to be less accurate than the freshman method, and they were considerably less accurate than the y-values found for the broadcast footage.