Week 8

Goals for this week include giving the Lego Midstorms presentation, add connected segments an convex hull information to my algorithm after first classifying the image as textured or non-textured, and continue investigating about computer science graduate programs, research topics in computer science. As a minor goal, update my website and make sure information remains current and functional.

The Lego Mindstorms workshop was last Monday, and it was an amazing experience! The seventh graders were a majority of girls. And they were very engaged with the programming of the robots. Though we had not anticipated it, one team built a triangle, another team for their final presentation had the robot go in a path that traced out two hearts, and a third team did both an hour glass shape and a five point star. At the end of the presentation students were asking if the Lego Mindstorms were for sale and where they could buy them. Some teams figured out how to do loops without being told. Some teams never figured out loops, they simply modified the time that a robot performed an action, so there were various levels of engagement and success. Many good ideas grew out of the presentation as well. In the end a 4^th person not trained on Lego Mindstorms helped out as well, he just wanted to try out the robots. All of us volunteering worked together to deliver a unified product, troubleshoot technical difficulties, and provide an experience the kids could take away that would spark their curiosity and about technology

--

 At the Tuesday’s meetings I showed my most current results obtained by training the classifier with an image set was very simple, containing either no texture at all (just precipitate) or crystals. For that data set I obtained a remarkable 10% false positive and 10% false negative rate. However, given the nature of the data set, we concluded that we cannot generalize to all images in general. So I must find a way to determine if the area of interest is textured or not.