In histogram shape-based watermarking schemes, watermark bits are embedded by altering the shape of a histogram extracted from the host image. Exited embedding algorithms use a group of histogram bins to embed only one watermark bit, which results in a rather low watermark capacity. In this paper, we improve the embedding algorithm in two new ways. The first proposed algorithm performs multi-round embedding to carry more watermark bits. In each round of embedding, a specified histogram is extracted so that the embedding operation does not affect watermark bits embedded in previous rounds. The second proposed algorithm uses a group of histogram bins to embed more than one watermark bits, where the coefficient transferring is optimized to minimize the embedding distortion. These algorithms can effectively enlarge the capacity. Furthermore, a histogram preadjustment method, together with a refined coefficient transferring method, is introduced. As a result, reasonable performances on robustness and watermarked image quality are available. The proposed algorithms provide various tradeoff among capacity, robustness, and perceptibility, which supports a wide range of applications.
CITATION STYLE
Feng, B., Weng, J., & Lu, W. (2017). Improved algorithms for robust histogram shape-based image watermarking. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10431 LNCS, pp. 275–289). Springer Verlag. https://doi.org/10.1007/978-3-319-64185-0_21
Mendeley helps you to discover research relevant for your work.