Rate-distortion based reversible watermarking for JPEG images with quality factors selection

The improved reversible data hiding scheme which is a part of JPEG coding process is introduced. Generally, one of the common constraints imposed on digital watermarking in frequency domain is a small payload that can be embedded without causing high degradation of a JPEG stego image. Moreover, even at small hidden payload the stego image file size will increase to some extent. In no existing data hiding technique compliant with JPEG there is a possibility to define in advance the file size of the watermarked image. Therefore, in this paper we propose to use a rate-distortion theory that minimizes coding distortion subject to a coding rate constraint. An iterative algorithm based on a Lagrangian formulation is applied to obtain a vector of quality factors for each of the 8 × 8 blocks that scale the JPEG standard quantization table. The experimental results show the advantage of the proposed improved watermarking scheme in terms of data payload versus quality and file size compared with the state-of-the-art data hiding schemes, and, furthermore, clarify the improvements of its optimized counterpart.