This dataset contains 50000 synthesized character images. Number of
examples in each class have been set so that it matches the
character frequncies optained using a large corpus.

Please note that the class label of the synthetic set is in the
range of 1-62, following the A-Za-z0-9 order, but the files
containing the ICDAR dataset uses ASCII values for the characters
as labels. If you use this data, please remember to set the
'whetherASCII' variable to false when calling the 
'extract1stLayerFeatures/prepData' function in our demo code.
