I plan to recognize paper bills by denomination but i can't think of a way to do so ... it must be invariant of illumination, rotation and scale. can i use haar or LBP ? do you have any suggestions?

edit

imho, training cascades on your problem is a bad idea.

cascades are binary classifiers (yes or no), training many of them on very similar things like banknotes will only result in many false positives.

i think, you're confusing a classification problem(which banknote is it?) with a detection one(banknote or not?) and cascade classifiers are the wrong tool for this.

so what should you prefer that i use that includes training?

Since you need to deal with illumination, rotation and scale, I suggest you can give bag of visual words(BOVW) or convolution neural network a try. opencv provide good tools to train your classifier base on BOVW, if you need a very easy to use cnn library, please google tiny-cnn.

Since you need to deal with illumination, rotation and scale, I suggest you give bag of visual words(BOVW) or convolution neural network a try. opencv provide good tools to train your classifier base on BOVW, if you need a very easy to use cnn library, please google tiny-cnn.

can i use bag of visual words for identifying the denomination of the paper bill ?? or can i use haar ? instead