complasas.blogg.se

With multiple fonts that are used in OCR are not addressed in this paper as they are Section, we survey the databases that are used for Arabic/Farsi FR. The lack of a benchmarkingĭatabase forAFRsystems resulted in researchers developing their owndatasets. 4, and nally theĭatabases are an important resource for FR systems. 3 Ĭlassication techniques and used features are presented in Sec. 2 preprocessing techniques used in AFR are presented in Sec. The remainder of this paper is organized as follows: databases for Arabic FR are In this paper, we have surveyed all Arabic and Farsi FR publications that we areĪware of, some of which targeted Arabic fonts, others targeted Farsi fonts, and someĪddressed the identication of both Arabic and Latin and Arabic and Farsi fonts. Nition rates and saving in processing times. Type is recognized, then a mono-font OCR may be used resulting in higher recog. May be an important preprocessing step in an OCR system. Variety of fonts renders the task of recognizing the font type a challenging task. There are more than 450 fonts available for Arabic and Farsi languages.41 This vast Moreover, Arabic characters may take dierent shapes based on the font. Language, most of its characters adopt several shapes based on their location within a Tion.4,40 Unfortunately, the same statement cannot be stated for Arabic FontĪrabic language consists of 28 characters. Languages like Malay, Farsi, and Urdu use Arabic characters for writing.4 RecentlyĬonsiderable improvements have been achieved in the area of Arabic text recogni. Moreover, some West African languages such as Hausa and non-Semitic Malaysia and Indonesia) that use it for reading and writing Arabic A cooperative approach combines a priori and a posterioriĪrabic language is used in Arabic countries in addition to the majority of IslamicĬountries (e.g. Text image, while the posteriori approach depends on the contents of the text to a priori, a posteriori, and cooperative or hybrid).15,37,50 In theĪ priori approach, the font is identied before recognizing the characters of the There are three approaches to combine Optical Font Recognition (OFR) and Time compared to human editing to recover the font and styles of text. The text but also the fonts and styles of the document. In addition, it enables the system to reproduce not only Ognition rates compared with Omni-font (which assumes a number of predeterminedįonts) and requires less time. In using Mono-FR system (which assumes one font type) that results in better rec. Recognizing the font before using Optical Character Recognition (OCR) helps #.c World Scientic Publishing CompanyDOI: 10.1142/S021800141553002XįR is useful in improving text recognition in terms of recognition accuracy and International Journal of Pattern Recognitionand Articial Intelligence Which is used in the classication phase to recognize the font styles.8,21 This representation contains discrete information Image-format conversion of the input image.8 The preprocessed image is then Phase applies de-noising, skew correction, normalizing, text segmentation, and Main phases: preprocessing, features extraction, and classication. Similar to other pattern recognition approaches, FR includes three The aim of FR is to recognize the font style based on features that are extractedįrom text images. (Bold, Italic, or Plain) of the printed text. Identifying the font style involves determining the font typeface, size, and weight Keywords : Arabic font recognition Farsi font recognition Arabic font databases classication įont Recognition (FR) is the task of recognizing the font style of a text image.47

The paper addresses the strengths and limitations of the presented techniques and specied areasof research that are not, so far, addressed in Arabic/Farsi FR as well as areas of possible To our knowledge, this is the rst survey of Arabic/Farsi FR and used databases. All published work of Arabic and Farsi FR, which the authors are aware of, are surveyed.

preprocessing, classication techniques and used fea. The mainphases of FR systems are surveyed (viz. Paper, we survey the literature of Arabic and Farsi FR research and used databases. InĪddition, it can be used to restore the original document text fonts, styles and sizes. **Qassim University, Qassim 51452, Saudi 13 December 2013įont Recognition (FR) is useful in improving optical text recognition accuracy and time. *King Fahd University of Petroleum & Minerals, Arabic and Farsi Font Recognition: Survey