So I woke up this morning and went over the H data. Part of me believes we can't do the project if they give it to us. When I look at the data while in depression we for sure can't do it. So today I woke up close to normal. I can feel my brain keeping the depression at bay. I think it's the prozac working. Any way, I look at several data images. In every case I was able to grab key features. (Finding a grabbing key features is an issue) Some much better then others and some barely (but I was able to grab some). And in some cases I got matches. Being honest with my self not in depression, this is how I feel based on my experience with this type of data. First, we need a much larger data set to work with. Second if we get a much larger data set and the same principles apply for OS matching with L??? We should be able to triangulate the abnormalities, based on cluster grouping and single cells. If we end up with too many false positive, we may be able to average the main color of the object add it to the key, and use that as a final filter. I realize if your reading this, it does not read very depressing to you. Part of my depression is caused by earlier failures of the technology. Since then I've added some improvements and fixed a bug or two. I also believe the failures I am alluding to were caused by inconsistent input format. I have run other test that have actually worked correctly. Well that's today's update.