Histogram Setup to recognize Arabic Calligraphy

摘要

Character recognition for any language has been under the perspective of active research for the past few years. With no exception, Arabic character recognition has also been therefor the past two decades. These two decades of work has shown active involved on character analysis using various techniques but yet the problem of character recognition is not solved due to the complexity of Arabic characters. In the face of the Calligraphy style of writing, the alphabetic representation is based on rules. So, in order to recognize these alphabets we can use those rules. In this paper, we specifically look at the problem of how to recognize the Arabic alphabets written in Thuluth style of calligraphy. Having considered the rules for the data representation, the data is analyzed using the projections, both vertically and horizontally. These horizontal and vertical projections are used as input to a neural network system to train the network and make it recognize the Arabic alphabets. In this paper, we cater the mode of recognition of Arabic alphabets written in Thuluth style using histogram. The alphabets are entered in a matrix format and the horizontal and vertical projections are got and given as input to neural network system to recognize the alphabets.