I found that if you label your samples to some kind of a pattern, the filtering with “contains” gives you just a little more insight to the data. As an example; if my database was fruits and veggies, I might make the first 3 letters define the type of fruit or vegetable(BAN-Bananna; GRP-Grape). The the 5th and 6th characters then are the subclass (APP RD; Red Delicious Apple;APP GS;Granny Smith Apple)..
The second is more of an enhancement:
It would be nice to be able to add or subtract scan groups from each other. That way if you keep getting that film of wax to make the apples look in your scans, you could scan that wax and remove it from all other scans.