Undetection of some characters

[Shweta Sharma]

Hi,

I am currently working on license plate recognition and for that i am using tesseract3.03 on windows. For the same purpose I have trained the tesseract with 1500 images of alpha numeric characters.When I am testing tesseract for same images then GetUTF8Text returns NULL.Specially mostly for N,M, and W characters.I have attached 3 images as samples. 

As detail, I have set the vaiable PageSegMode="PSM_SINGLE_CHAR".Besides,I have already tried preprocessing and resolution techniques for input image to resolve the problem.

PLEASE HELP!!!

It would be helpful even if I can know the roll of GetUTF8Text in giving the NULL output for an images.