Two-dimensional difference expansion (2D-DE) scheme with a characteristics-based threshold

Fragile watermarking is commonly used for content authentication and tamper detection in digital multimedia. For some critical applications, such as medical systems, the fragile watermarking system should be based on a reversible data hiding scheme. The Difference Expansion (DE) technique is a popular reversible data hiding method. In this paper, we adopt a two-dimensional difference expansion technique (2D-DE) to increase the hiding capacity. The proposed scheme is a fragile block-based scheme where the image is divided into blocks of 4×4 pixels, and these expandable blocks are embedded with 16 bits of payload. In addition to increasing the achieved capacity, we study the effect of selecting the blocks to be embedded according to characteristics such as standard deviation, smoothness, and uniformity. These characteristics can be used as selection criteria to enhance the visual quality. The experimental results demonstrate a very high embedding capacity equals to approximately 1 bit per pixel (bpp). Moreover, a threshold based on the standard deviation can be used to control the hiding capacity and predict the reduction in hiding capacity that will enhance the visual quality.

► We study the relation between image characteristics and MSE after embedding data.
► We propose a threshold based on standard deviation to control data hiding capacity.
► The proposed threshold enhances the visual quality of the embedded image.