Abstract—Automatic recognition of handwritten numerals has been widely proposed in various languages. However, some languages such as Persian still need more consideration. In this paper, we proposed a handwritten Persian numerals dataset, which is gathered from people with different range of educational level. Thus, it is more general than other similar Persian datasets. Additionally, a method to classify Persian handwritten numerals is presented, which uses simple geometrical features based on their shapes, and classifies them via a rule-based decision tree classifier. Compared to other similar methods, our proposed method has the advantages of high speed running, employing too few and simple features, and elimination of training phase.
Index Terms—Decision tree, feature extraction, geometrical shape, preprocess.
The authors are with the Shiraz University, Shiraz, Iran. (e-mails :{alvari,hazrati}@cse.shirazu.ac.ir, baharsalehi@gmail.com ).
[PDF]
Cite: Hamidreza Alvari, Seyed Mehdi Hazrati Fard, and Bahar Salehi, "A Decision Tree Based Method to Classify Persian Handwritten Numerals by Extracting Some Simple Geometrical Features,"
International Journal of Computer Theory and Engineering vol. 5, no. 1, pp. 118-122, 2013.
