Need help on OCR binarization and segmentation of character

I am trying to 'train' my OCR software to recognize numbers in the Bengali language. I have the code set up to binarize and segment loaded samples of numbers. But when I try to use it to train the software, I get this error message:

Too many outputs requested.  Most likely cause is missing [] around
left hand side that has a comma separated list expansion.

Error in ==> train at 229
    [I.label,num] = bwlabel(I.bw);

??? Error while evaluating uicontrol Callback

I am using Matlab 2008a

Answer 1

[Walter Roberson]

Is "I" a structure array? If so, then I.label would expand to a list of names, as if you had written

[I(1).label, I(2).label, I(3).label, ...,I(end).label, num] = bwlabel(I(1).bw, I(2).bw, I(3).bw, ..., I(end).bw)

Answer 2

[Rim]

Thanks for your reply.Could u please go through the code and figure out wht is the problem.I can't fix it.

the code is given in this link.

http://www.mediafire.com/?ck5culxpka2d1g6

[Walter Roberson]

Sorry, I don't open .rar files.

Answer 3

[Rim]

ok i have added the code in pastie.org here is the link for train file: http://www.pastie.org/1811553 find_means: http://www.pastie.org/1811566

training image is uploaded on the given link: http://img831.imageshack.us/i/oneline.png/

[Walter Roberson]

Sorry, I do not have the toolbox that would be needed to test that code myself.

At the command line, give the command

dbstop if error

Then run your program. When the error occurs you will be put in to the debugger. Examine the classes and sizes of the variables involved -- e.g., determine whether I is a structure array at that point. If so then we will have identified the most immediate cause, but it might take some work to track that back to particular input conditions.