A Permutation-based Reversible Data Hiding Method with Zero Visual Distortion

Recently, several approaches have been designed to hide data in portable document format (PDF) files. These approaches have demonstrated their advantages in different application scenarios, including copyright verification, covert communication/steganography, and content forensics. However, they often suffer from visual distortion or lack universal applicability. In this work, we propose a reversible and transparent method that exploits the coding properties of text objects (i.e., substrings) in a PDF-compliant document to embed data. In particular, the position information of the substrings is adjusted to hide data, where each unique permutation of the substrings encodes a bit sequence. Subsequently, the distance of each substring from the left margin is corrected so that the processed PDF has the exact layout or appearance of the original PDF, hence completely preserving the quality of the original PDF file. In the best-case scenario, to hide one bit of data, 5.88 bits of the PDF file are required, i.e., 1 : 5.88. In addition, this method can be deployed in tandem with conventional data hiding methods to hide more data and to hide data in different ways.