On the determination of capacity parameters in PEE-based reversible image watermarking

In the existing prediction-error expansion (PEE)-based reversible image watermarking schemes, the capacity parameters are determined in a recursive manner by gradually turning these parameters to fit the payload. This class of method needs multiple rounds of embedding iterations, and hence, it is computationally inefficient. In addition, when multiple capacity parameters need to be handled, the previous methods are generally not capacity-distortion optimized. In this work, we formulate the task of determining the capacity parameters as a capacity-distortion optimization problem, which can be shown to be convex. We also prove that under some conditions, even simple analytical solutions exist.