中文摘要：

为了解决目前基于块差值无损鲁棒水印算法的不足,提出了一种基于Huffman编码和K-means聚类的改进算法。对嵌入过程中产生的水印负载,利用Huffman编码进一步减少其大小,提高水印嵌入容量;对水印提取过程中,可能发生的1-bit区域和0-bit区域重合的问题,利用K-means聚类算法提高水印提取精确度。实验结果表明,与现有相关算法相比,文种算法在水印嵌入容量和鲁棒性上具有明显优势。

英文摘要：

To deal with the problems in lossless robust watermarking based on image block~ s difference, an adaptive algorithm is proposed. In the embedding procedure, the Huffman coding is employed to reduce the watermark＇ s overhead such as the marked information of image＇s block, and increases the actual embedding capacity. In the extract procedure, the K-means clustering is utilized to resolve the overlap problem between 1-bit zone and O-bit zone, and enhances the extracting accuracy. Experimental results show that, compared with previous works, the performance of the proposed method is significantly improved in terms of capacity and robustness.