Meeting Date: 30.08.1999
What I did last week: I've basically been a good boy in that I did all the things that I planned. I restructured the object-layout, so that the VRGestureRecognizer became a VRVotingRecognizer, the VRGestureRecognizer became a base class for the VotingRecognizer and the VRSizeFunctionRecognizer. The VRGestureActor now has a private variable of type VRGestureRecognizer and instantiates one of the Recognizers depending on what kind of gesture data he finds in the directory given in the constructor. The SizeFunctin one takes priority. Also I made a training application for the SF-Recognizer, so that one can compile trainingsets. I also included the option of generating trainingdata for both Recognizers from the constructor. I then proceeded to re-enter the data for the files containing the angular data for each gesture (*.ang files) using the polhemus tracker (which for me works even better than the touchpad... unfortunately not for everyone ;-(
I also changed the format of the trainingset-files to contain the names of gestures to be recognized. For one thing it makes it easier for people to understand trainingset files (I sure wish I would've had that kind of clue!) and also it allows the automation of gesture-loading by the SF-Recognizer. Then I did a whole lot of Prac work and actually started on the layout of my Write-up (structuring the chapters). I will ask Shaun what he thinks about it and then get going...
What I'll do this week: More Prac work and three lectures a day will keep me pretty busy. I still want to get some user-response data for my project and am thinking about what and how to get at the moment. In my "spare"-time (whatever!) I will start on my write-up and somewhere between the 25th and 27th hour of the day I will breathe. Please stay tuned and I will see you again next week ;-)