Abstract
The reversible watermarking algorithm based on pixel value ordering (PVO) embeds secret data by modifying the maximum and minimum values in a pixel block. The performance of this algorithm heavily depends on the inherent correlation among adjacent pixels within an image block. To improve the inherent correlation among adjacent pixels within an image block, a new method to incorporate the dynamic block-partition strategy into the PVO algorithm is proposed in this paper. This new method includes the following procedure: First, the host image is divided into non-overlapping regions according to the image pixel values; then, each region is partitioned into different blocks, which are subsequently classified according to the local complexity and predefined threshold values for embedding; and finally, watermark embedding is performed using the PVO-based algorithm. In the proposed method, the pixel values of each embedded block are located in a relatively small region to improve the inherent correlation and thereby enhance the embedding performance of PVO. The experimental results show that the proposed algorithm has a better embedding performance compared with that of the conventional PVO based algorithm.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call