In the End, Research Becomes DDR

What a hectic week. As I am trying to complete the end of my summer program, I am also trying to run a bunch of school errands and putting my house mate on my lease. Indeed, this week is crazy.

Monday was the mocap day. I went into the motion capture lab prepare to do 50 reps of an exercise correctly and incorrectly. I didn't make that mark as I only did 10 reps each of the exercise. That was know easy fit as each rep consistent of 10 reps so I was doing 100 individual standing hamstring curls with 10 pound cuff weights. I feel like I am going to have to do motion capture of the exercise again and mentally prepare myself to do 50 reps (500 individual curls) each of correct and incorrect motion to actually have a valid training set. That is a lot of data!!!!

Tuesday, I met with Ron and Jessica as usual. We discussed my progress. It was decided I would finish my code that would convert the motion capture data into file format that would be read by WEKA and run a machine learning algorithm with WEKA on the data. However, with a test file that contain a motion capture file I ran OUT OF MEMORY!! Of course, that always happens with Java because its small heap size. Note, I am looking at a file that would contain data from 50 plus motion capture files. I will have to rethink how I will analyze this data for some initial results. Looks like I am heading back to Torch and Adam's code and resaerch project for answers.

Besides doing research, I have been using my break time to master DDR (Dance Dance Revolution). In addition to research and break time, I have perusing dancers and dance styles to motion capture for CMU public motion capture database.