Imad I. Methods of processing texts in Arabic alphabet

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0408U004730

Applicant for

Specialization

  • 05.13.06 - Інформаційні технології

30-09-2008

Specialized Academic Board

Д 35.052.14

Lviv Polytechnic National University

Essay

This dissertation is dedicated to the development of information technology for processing distorted images of the text in Arabic language to improve their automatic recognition. It has been demonstrated, that at obtaining the initial image or in case of poor quality of the original (for example, from a photograph, fax message page, etc.) there can be distortions of the text in the form of poor sharpness or smears. . The work presents characteristics of the texts in Arabic language, shows complexity of the problems of recognition of such texts caused by the presence of diacritic marks and different manners of writing the letters depend-ing on their place in the word (in the beginning, in the middle or in the end) or in the sentence sepa-rately, by difference in the style of writing, and pro-poses a method of recognition of the signs that join together or become imposed on each other. Mathematical problems of preliminary processing of distorted by smearing or unsharpness images have been considered to resolve the problem of restora-tion. It has been shown that the problem of recon-struction of a distorted image boils down to the solving of appropriate first-order Fredholm equation solved by the method of regularization according to the suggested formulas

Files

Similar theses