A new algorithm for machine printed Arabic character segmentation

Authors:
Liying Zheng;Abbas H. Hassin;Xianglong Tang
Affiliations:
School of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China and School of Computer Science and Technology, Harbin Engineering University, Harbin, Heilongjiang, ...;School of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China;School of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China
Venue:
Pattern Recognition Letters
Year:
2004

Citing 4
Cited 4

Knowledge-based English cursive script segmentation

Pattern Recognition Letters
A new segmentation technique for omnifont Farsi text

Pattern Recognition Letters
Off Line Arabic Character Recognition - A Survey

ICDAR '97 Proceedings of the 4th International Conference on Document Analysis and Recognition
Off-Line Handwritten Arabic Character Segmentation Algorithm: ACSA

IWFHR '02 Proceedings of the Eighth International Workshop on Frontiers in Handwriting Recognition (IWFHR'02)

Fast Zernike wavelet moments for Farsi character recognition

Image and Vision Computing
A pictorial dictionary for printed Farsi subwords

Pattern Recognition Letters
Semi-continuous HMMs with explicit state duration for unconstrained Arabic word modeling and recognition

Pattern Recognition Letters
Segmentation of Arabic Characters: A Comprehensive Survey

International Journal of Technology Diffusion

Quantified Score

Hi-index	0.10

Visualization

Abstract

The major problem with machine printed Arabic character segmentation is the shape of the letter depending on its location in the word. In this paper, a new machine printed Arabic character segmentation algorithm, which is based on the vertical histogram and some rules, is presented. The rules which are based on, not only the structural characteristics between background regions and character components but also the characteristics of isolated Arabic characters, are used to check whether the sub-word includes only one character. Then we use the vertical histogram and some other rules to find real segmentation points. Finally, we split the sub-word at the segmentation points. The experimental results show that the algorithm achieved about 94% correct segmentation.