A new large-scale multi-purpose handwritten farsi database

Abstract

This paper introduces the Center for Pattern Recognition and Machine Intelligence (CENPARMI) Farsi dataset which can be used to measure the performance of handwritten recognition and word spotting systems. This dataset is unique in terms of its large number of gray and binary images (432,357 each) consisting of dates, words, isolated letters, isolated digits, numeral strings, special symbols, and documents. The data was collected from 400 native Farsi writers. The selection of Farsi words has been based on their high frequency in financial documents. The dataset is divided into grouped and ungrouped subsets which will give the user the flexibility of whether or not to use CENPARMI's pre-divided dataset (60% of the images are used as the Training set, 20% of the images as the Validation set, and the rest as the Testing set). Finally, experiments have been conducted on the Farsi isolated digits with a recognition rate of 96.85%. © 2009 Springer Berlin Heidelberg.