摘要
针对一类单幅图像的复制遮盖篡改,提出一种基于快速傅立叶变换相位和幅度的自相关检测算法对篡改部位进行检测。利用Canny算子对篡改图像进行边缘提取,对提取的边缘图像进行分块傅立叶变换,比较块与块之间的相似性程度确认图像中被篡改的部位。实验结果表明该算法能够有效检测出图像被篡改部位,并且对图像复制部位的润饰处理和一般的噪声攻击具有鲁棒性。
To detect the forged parts in a single image with copy-move attack,an autocorrelation detection method based on phase and amplitude spectrum of FFT is proposed.Firstly,the edge of the tampered image is extracted using the canny operator.Then,the edge image is processed with block-based FFT.Finally,the similarity between blocks is calculated to locate the position of the tampered area.Experimental results indicate that the method can detect the forged parts effectively,and is robust to general attacks and retouching of the copied area.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第9期2121-2123,2126,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(60772122)
安徽省高校省级自然科学重点研究项目(KJ2008A033)
关键词
快速傅立叶变换
被动认证
自相关检测
边缘提取
篡改定位
fast Fourier transform
passive authentication
autocorrelation detection
edge extraction
tamper localization