I am working on recognition of faces in given videos with provided Data base.I am using createFisherFaceRecognizer() . But I am not getting Proper output. is it problem with Data base or the predict method, which i have used.

edit

this is mostly a problem with your data. possible reasons:

• not enough images per person (20+ needed)
• images too large or too small (try 100x100 or such)
• bad cropping (too much background)
• bad alignment (not centered, not straight forward)
• missing lighting normalization
• different processing for training & prediction

setting up the code is easy, tweaking your data to get a good performance is hard here !