The separated RGB character images of 1 00 by 1 00 pixel
resolution obtained in the fist part are converted into
grayscale images and then the grayscale images is again
converted into binary images using appropriate gray scale
thresholding. Now, binary image is thinned using
skeletonization infiite times. Edges of these thinned images
are detected using appropriate thresholding and then further
dilated using appropriate structure element [ 1 9,20] . These steps are known as preprocessing. The preprocessing steps are
given in the following fiure.