Prediction-Based Reversible Data Hiding with Content Characteristics

Reversible data hiding is one of the popular topics in watermarking researches, and it belongs to the branch of digital rights management (DRM) applications. Similar to conventional watermarking techniques, for reversible data hiding, at the encoder, secret information can be embedded into original images, and the marked image can be obtained. Different from conventional watermarking, due to the term `reversible', it implies that at the decoder, both the original image and embedded secret should be perfectly separated from marked image. It requires keeping the reversibility of proposed algorithm, while looking for good image quality and large amount of secret for embedding. With the prediction-based algorithm, output image can be predicted, and differences between original and predicted images can be altered to make reversible data hiding possible. We also utilize inherent characteristics of original images for reaching better performances. Simulation results reveal that with our algorithm, comparable or better performances can be observed due to the characteristics from different images.