中文摘要：

为了提升数字图像内容认证的准确性,提出一种新颖的半脆弱水印算法.引入相关块组的概念,并将相关块组进一步划分为子块组,轮流使用每个子块组内特定块生成水印,最终将水印嵌入到子块组内的其他块中.算法根据图像内容自适应选取量化步长,结合Slant变换高能量集中性及直流量化原理完成水印嵌入过程,认证过程中使用新的噪声过滤策略并通过投票方式确定篡改发生位置.实验表明：该算法能有效地提高含水印图像的质量及防止水印信息的未授权提取,在应对裁剪、替换等常规处理时定位精度提升明显.算法抵抗偶然攻击的能力强,新噪声策略有效地改善了针对篡改区域的识别效果.

英文摘要：

A novel semi-fragile watermarking algorithm was proposed to improve the accuracy of image content authentication.This algorithm introduced the conception of relevant-block-group,and divided each relevant-block-group into several sub-block-groups,generated the watermark information by using the particular block of the each sub-block-group by turn,finally,embedded the watermark into the other blocks of the sub-block-group.This algorithm selected the watermark quantization steps self-adaptively according to the content feature of the host image,finished the embedding process by combining the high energy concentration of the Slant transform and the principal of direct current coefficient quantization.Moreover,a noble noise-filtering strategy was proposed,and the voting method was applied to position the tampered area in the authentication process.Experimental results show that the proposed algorithm can improve the quality of the watermarked image effectively and prevent the watermark detection without authentication,and it can also significantly enhance the positioning accuracy when facing with conventional treatments,such as cutting,replacing,and etc.The algorithm has good ability of resisting accidental attacks,and the new noise strategy improves the recognition effect of the tampered region effectively.