Text Steganography with High Embedding Capacity Using Arabic Calligraphy

Abstract

Text is a challenging media for steganography due to the limited amount of redundancy bits in text and the ease of detecting minor changes in it. In linguistic steganography, the features of a language are used to hide data. The more features we have in a language, the more capacity of hiding data we can get. The Arabic language is rich in its features, many of which are still not used in text steganography. For example, Arabic calligraphy includes a variety of fonts and shapes for the Arabic alphabet. In this paper, by modifying a string matching algorithm and using suffix tree, the generation of Arabic stego text according to a secret message is implemented. The generation process is performed using a linguistic framework that utilizes Arabic calligraphy (AC-STEGA). In the presented work, multiple shapes of letters in Arabic font (Naskh) are used to increase the capacity of hiding data. This work comes as an extension of previous work to apply one shape of letters in (AC-STEGA). The evaluation showed an improvement in capacity of 65.4% compared to 25.4% presented in previous work. To satisfy security requirement, we did not use cover, the cover represents the secret message itself.