An adaptive prediction-error expansion oriented reversible information hiding scheme

Data hiding in digital images can be used in secure communication, copyright protection, and etc. For some important images, such as medical and military images, the original images must be recovered after extracting the embedded data, because distortions are unacceptable for these kinds of images. In this paper, we propose a reversible data hiding method based on prediction-error expansion. Each pixel of the cover image, excluding the first row and the first column, is predicted by its top and left neighboring pixels in the raster-scanning order. The relationship between the prediction error and the pre-determined threshold decides whether the current pixel is embeddable or not. Since the proposed prediction process provides small prediction error, our method can achieve high embedding rate and good visual quality of the stego image by the expansion of prediction error. During the procedure of extraction and recovery, the same prediction process is conducted, and then the embedded secret data and the cover image can be recovered correctly. The histogram squeezing technique is utilized to prevent underflow and overflow problems. Experimental results show that the proposed method provides better performance than some other methods.

► Prediction-error expansion is used to embed secret data in cover images reversibly. ► Current pixel is predicted by its top and left pixels in raster-scanning order. ► Higher embedding rate and image quality are achieved due to small prediction-error. ► Underflow and overflow problems are prevented by histogram squeezing technique.