hi all..
i am working on my final year project,,its about CAPTCHA breaking...i.e. to show that captchas can be easily breakable....
for that following steps need to be followed:
1. firstly trace the position of characters in the input image.
2. then extract each character,,we can also say it as character segmentation..i.e. separating each character individually...
<<i have completed my project till dis step>>

now we have :
1. the segmented images...
2. a training set of images i.e. a set of images each representing an individual character.

what i want :
1. maintain a database in which 2 columns will be there... one representing the individual character i.e. from a to z...and second representing the value obtained after applying some function on the corresponding training image.
2. now after applying the same function on the segmented image which we have obtained in step 2,,we'll get their values also...now these values will be compared in the database,,,nd the corresponding character'll be fetched.....
in this way v can get the text corresponding to that image...
what say guys.......
i just wanted to know which function should i apply on the training set of images..........
..plz guys help me in this work... plz....