Reversible Data Hiding for Difference Alteration with Authentication Capabilities

Reversible data hiding is a newly developed research field in digital rights management (DRM) applications. It is a branch in watermarking researches, with the characteristics of reversibility at the decoder. Secret information can first be embedded into original image, with algorithms designed by researchers, at the encoder. Next, image containing secret data is delivered to the decoder. After reception of necessary side information for decoding, original image and secret information need to be perfectly recovered at the decoder. Based on the characteristics of original images, by carefully modifying difference values between neighboring pixels, the larger amount of secret data can be embedded. In addition, the starting point can be randomly chosen, and the location of such a point and the embedding strength can be encrypted for authentication at the decoder. We propose an algorithm that is able to hide more capacity with the reasonable output image quality, and to retain the capability of authentication to look for better security of the algorithm. Simulation results have demonstrated the better results over existing ones. Authentication at the decoder can also be validated.