Intelligent Feature Extract System for Cursive-Script Recognition

Abstract

Summary. The paper describes a newly presented hybrid method for high efficiency in script image feature extraction. The recognition rate was about 82% for very large number of scripts per class. However, it has reached even 100% in some cases with a smaller number of scripts per class. The system contains two projectionbased methods for image characteristics extraction presented by very simple feature vectors and one image descriptor. A specially worked out thinning algorithm for the recognition system has simplified the feature extracting procedure as it provides a continuous one-pixel width skeleton of the script, which is essential for the simple-projection approach.

Intelligent system for Arabic character recognition

“…x The rate of recognition is 93.5% when the input is the rest of Arabic characters of Arial and Courier fonts of size (10,12,14), which are not picked up in training. The NN discriminates 72 letters out of 77 input characters.…”

“…On the basis of the basic characters of 16 Arabic letters in Arial and Courier fonts of multi size (10,12,14), the NN training is implemented at workout time of 29…”

“…The NN is applied to 173 Arabic characters, Courier and Arial fonts of multi size (10,12,14), and is recalled for 173 letters. Table 8, and the overall recognition rate is 98.7 % for this system.…”

“…At training time of 55 [seconds] with 689 [cycles], NN is trained on 28 characters of Courier font of multi size (10,12,14). The rate of recognition is 100% through the recall process of 85 characters.…”

“…That is why, the work on Arabic characters recognition was limited. Efficient works on recognition of cursive-character script using Toeplitz model are presented in [6], [8], [10]. Other works have dealt with Typewritten Arabic characters using Zernike moments as feature extractor [1] or shape matching for character recognition [5].…”