python - Classify collection of images -

i have folder collection of images microscope , have separate them 2 classes (samples defects , without defects). additionally i've got sets of classified images. never tried before have example of how using python scikit library?

not question here, since there's programmatic side, i'd try help.

this 1 solution, mind you.

the problem breaks down to:

1. in these kinds of problems, first thing need figure out "features" of photos distinguish two. example,maybe "good" class has more curved lines / circles in it, , "bad" class has mor linear/sharp lines. or maybe 1 class more light , other dark, etc.
2. the second part, each such "feature" create method score value image. each photo value (say between 0.0-1.0, not necessary), , each photo has feature vector.
3. using inputs of feature vector each image in training set, can train decision tree. @ http://scikit-learn.org/stable/modules/tree.html, helped me.
4. once have classifier ready, apply here.
5. mind you, whole art here creating right "features".

alternatively, might want ath violla-jones image classifier, can use opencv train this. 1. explanation of how train classifier: http://docs.opencv.org/trunk/doc/user_guide/ug_traincascade.html 2. paper explaining it: http://www.cs.cmu.edu/~efros/courses/lbmv07/papers/viola-cvpr-01.pdf 3. tutorial http://note.sonots.com/scisoftware/haartraining.html

hope helps