The embedding of Arabic characters in QR code

Abstract

Since the introduction of QR Codes in the 90's, many companies have developed QR Code generators and mobile applications to read QR Codes. Despite the obvious benefits of next-generation data transfers, they have yet to be widely accepted and practiced. One of the main reasons is the use of unique language characters such as Arabic characters, Japanese Kanji, Thai Abugida and so forth. An attempt at embedding Arabic characters into QR Code for example, will result in the character's Unicode representation to be automatically encoded and/or decoded by default, instead of the original alphabet as a result of a method called Romanization. Thus, the intended use of the generated code has been useless. It is also noted that Romanization causes each character to be encoded in its own unique codeword which devours more space than encoding it with codeword for character combinations. In order to help solve this problem, this paper proposes the use of an algorithm that compares each word from an Arabic input with an internal codeword library that will then select the shortest codeword possible for single character and two-character combinations to replace the default Unicode representation, which also serves as a space saving method that allows longer messages to be encoded in a single QR Code. In return, the same method is also used to properly decode the message.