Arabic Digits Recognition Using Statistical Analysis for End/Conjunction Points and Fuzzy Logic for Pattern Recognition Techniques
Authors: Majdi Salameh
Number of views: 131
Arabic Digits recognition has the lights spotted on lately, since it could be useful in a wild range of fields. This paper provides an easy and fast technique to recognizing Arabic digits. This paper presents two methods about enhancing recognition rate for typewritten Arabic digits (Hindi). First, is node method that calculates number of ends of the given shape and conjunction nodes as well, the second method is fuzzy logic for pattern recognition that studies each shape from the shape, and then classifies it into the numbers categories. Two stages are going to be done by the two given methods, to recognize the Arabic digit, each come out with its own result and afterward compounds these result to obtain the final solution and statistical analysis. Several steps are taken in the recognition system, starting with the image processing, then feature extraction and the last step is classification. The image processing includes converting into binary, cropping the digit in single image, and getting a skeleton of the shape by thinning it. Feature extraction includes number of terminal and conjunction nodes from nodes method and two characters to specify the curve lines group for shapes and third number to know the position of end nodes according to conjunction nodes in similar digit such as ٧ , ٨ . The recognition includes compound between two vectors, one from each method. The proposed technique was implemented and tested the experimental results give high recognition rate for some fonts and either less for other fonts because of due to irregularity of some fonts (Andalus) or failing for one of the methods. The dataset contains multi-size for the digits from ٠ to ٩.