Adaptive reversible data hiding with pyramidal structure

Abstract

In this paper, we propose an adaptive algorithm for reversible data hiding by employing the characteristics and pyramidal relationships of original images. The major goal of reversible data hiding is to keep the reversibility of algorithm. By use of the pyramidal structure to explore the inherent characteristics of original images, regions with different smoothness levels can be determined, and then data hiding can be performed adaptively with the pre-determined threshold for balancing the output image quality and embedding capacity. On the one hand, larger capacity can be hidden into smoother regions with limited degradation of output image quality. On the other hand, the size of location map, which serves as the side information for keeping reversibility, can be reduced for embedding into smoother or less smooth regions of original image. By carefully manipulating difference values between layers in pyramidal structure, secret information can effectively be embedded. With our method, we observe better performances over relating methods with enhanced image quality, the more embedding capacity, and comparable amount of side information for decoding. More importantly, the reversibility of our method is guaranteed, meaning that original image and secret information can both be perfectly recovered at the decoder. Simulation results demonstrate that proposed method in this paper outperforms those in conventional algorithms.