User:TheLarry/Notebook/Larrys Notebook/2009/11/06

Tracking Software Upgrade

Main project page Previous entry      Next entry

Pattern Matching So I am looking at the folder of images that Andy gave me and a problem like i mentioned yesterday is that sometimes the thresholding cuts the μT in half or thirds or whatever because there is a spot missing because of low fluorescencing. So to fix this i thought of using a pattern matching where the microtubule from the frame before would serve as my template. I ran a quick check on this to see if it'll work and it looks like it will. Pretty nicely actually. But the problem is to implement it fully i have to add a couple of case structures. So the condition is if after the magic wand if the object has a much lower area then try the pattern matching algorithm. But my code is extremely messy (embarrassingly so) so making sure all the correct wires run through the right case structure can be annoying to do. I think i can do it and i think it'll help a lot so i'll hafta roll up my sleeves and try it out. It looks like i have the thing implemented well but i have a FUCKING RACE CONDITION somewhere this is fucking my shit up. I am gonna find it. OK i definitely fixed the race condition and i was mighty stupid to have made that mistake. Anyways, another bug came up and i think i fixed that one, but now that breakpoint that surrounds the entire screen occurred. and i am too tired from bugs to deal with it. so i'll handle it tomorrow. i think everything works now which is a bonus too