Yoruba Language and Numerals’ Offline Interpreter Using Morphological and Template Matching

Abstract

Yoruba as a language has passed through generation reformations making some of the old documents in the archive to be unreadable by the present generation readers. Apart from this, some Yoruba writers usually mixed English numerals while writing due to brevity and conciseness of English numeral compare to Yoruba numerals which are combination of several characters. Re-typing such historical documents may be time consuming, therefore a need for an efficient Optical Character Reader (OCR) which will not only  effectively recognize Yoruba texts but also converts all the English numerals in the document to Yoruba numerals.Several Optical Character Reader (OCR) systems had been developed to recognize characters or texts of some languages such as English, Arabic, Japanese, Chinese, and Korean, however, despite the significant contribution of Yoruba language to historical documentation and communication, it was observed that there is no particular OCR system for the language. In this paper correlation and template matching techniques were used to develop an OCR for the recognition of Yoruba based texts and convert English numerals in the document to Yoruba numerals. Experimental results show the relatively high accuracy of the developed OCR when it was tested on all size Yoruba alphabets and numerals.