The Segmentation of Printed Arabic Characters Based on Interest Point

Abstract

Arabic characters are different compared to the other characters whether from their forms or the way they are read. Before conducting a recognition process, we should conduct segmentation or divide each character to identify each Arabic character of the word. The enormous problem of segmenting the connected Arabic characters is dividing each character with different positions, forms, and sizes for each character. Therefore, we suggested a method in segmentation process by using the interesting point, which successfully obtains the 86.5% average accuracy.