Authors : Esmail Miri, Seyyed Mohammad Razavi and Naser Mehrshad
Abstract: In this study, a simple way to recognize the Persian typed sub-words are provided. First, the search space is limited by using a few simple characteristics and according to the position of points and signs of input sub-words. Then by using the ofloci features, two sub-words with the smallest distance than input sub-word are selected. Also, two sub-words are selected by using zoning features and minimum distance criteria. To select final sub-word among following four sub-words, the number of text vertical intersection with field feature is used. This method was tested for Lotus font and 98.18% recognition rate is achieved for test data.
Esmail Miri, Seyyed Mohammad Razavi and Naser Mehrshad, 2017. Recognition of the Persian Typed Sub-Words with a Hierarchical Manner. Journal of Engineering and Applied Sciences, 12: 2009-2017.