LSB based steganography with OCR: an intelligent amalgamation

Steganography, the art of hiding information within information, has an added advantage over cryptography as the person viewing the object in which the information is hidden has no knowledge of the presence of any hidden information. The Least Significant Bit (LSB) embedding technique which is one of the widely used techniques ensures an indiscernible change in the cover image oblivious to the human eye. In this paper, an Optical Character Recognition (OCR) based Steganographic technique is introduced, in which message, in its feature form, is embedded in the cover image. We extract character level features from images which contain the textual message, and embed these features in the cover image, strengthening the data hiding objective of steganography. This is because an intruder has to know about the presence of features as hidden bits and even thereafter has to have trained OCR model to retrieve the texts from the decoded information i.e. from the features. We validate our results on an English Printed Character dataset (Chars74K Dataset), and present the evaluation results for varied LSBs.