| 基于模糊不变矩和区域生长的图像盲检测算法 |
| |
| 引用本文： | 刘潘梅,吴建源.基于模糊不变矩和区域生长的图像盲检测算法[J].贵州教育学院学报,2011(6):14-17. |
| |
| 作者姓名： | 刘潘梅 吴建源 |
| |
| 作者单位： | 广东培正学院计算机科学与工程系,广东广州510830 |
| |
| 摘 要： | 提出了基于模糊不变矩和区域生长的篡改盲检测算法。该算法根据图像模糊原理,分析复制-粘贴篡改区域的特点,采用24个模糊不变矩表示图像子块特征;通过子块特征的相似度确定篡改区域的侯选种子,再根据侯选种子的邻域块对的相似度找出篡改区域的真种子,并通过中心连接区域生长法从真种子中生长出经过模糊的篡改区域。本算法对粘贴区域被模糊使之融入到背景之中的篡改图有较好的检测性能。
|
| 关 键 词： | 模糊不变矩 区域生长 图像篡改 盲检查 |
Image passive-blind detection based on blurred invariant and region growing |
| |
| LIU Pan-mei,WU Jian-yuan.Image passive-blind detection based on blurred invariant and region growing[J].Journal of Guizhou Educational College(Social Science Edition),2011(6):14-17. |
| |
| Authors: | LIU Pan-mei WU Jian-yuan |
| |
| Institution: | (Department of Computer Science and Engineering,Guangdong Peizheng College,Guangzhou,Guangdong,510830 China) |
| |
| Abstract: | An image authentication algorithm based on blurred-invariant and region growing is proposed.The algorithm uses 24 blurred-invariants as the image block＇s feature according the image blurring theory and invariant theory by analyzing the characteristics of duplicate-paste region;and finds out the tampered area candidate seeds through the similar degrees of block features,then determines the true seeds according to the similar degree of the neighborhood of the candidate seeds,and utilizes the center-connected region growing to grow the tampered and blurred area from the true seeds. |
| |
| Keywords: | blurred-invariant region growing image forgery passive-blind detection |
| 本文献已被 维普 等数据库收录！ |
