I don't know how the Labview Vision system works, but I would strongly suggest looking into the use of neural networks in your system. Neural network algorithms have been proven to be more effective in analyzing PET images than traditional ones. There have been a lot of graduate thesis done on the subject so the technical papers wouldn't be too hard to find. It's the technology american military intellegence uses to analyze PET and satellite images quickly and effectively, but unfortunately, they don't like to share their algorithms :).
Just look up the use of neural networks in terrain classification.
---------------------------------
[will work for kudos]
