基于高斯金字塔分解和圆谐傅里叶矩的图像复制移动检测算法
The image copy-move detection algorithm based on Gaussian pyramid decomposition and circular harmonic Fourier moments
-
摘要: 针对图像子块滑窗方式的移动导致数字图像复制移动检测算法的空间复杂度和时间复杂度很高的问题,提出基于高斯金字塔分解和圆谐傅里叶矩的检测算法.该算法首先对图像进行高斯金字塔分解,提取图像的低频区域,然后计算每个图像子块的3阶圆谐矩,并对各个图像块的特征向量进行字典排序,最后根据阈值进一步找到图像的复制粘贴区域.实验结果表明,该算法对经噪声污染和重压缩后的篡改图像具有很好的检测效果,且运算量较小.Abstract: Aiming at the problem that the space and time complexity of digital image copy-move detection algorithm are very high because of the movement of image sub-block sliding window approach,the detection algorithm based on Gaussian pyramid decomposition and circular harmonic Fourier moments was proposed.The algorithm firstly finished the Gaussian pyramid decomposition of the image and extracted the low frequency component of image,then calculated 3-circular harmonic Fourier moments of each image sub-block and completed the dictionary sort to feature vector of each image sub-block,finally further find the image copy-taste area according to various threshold.The experimental results showed that this algorithm had good detection effects for the tampered image added noise or compression again,and less compution.
计量
- PDF下载量: 29
- 文章访问数: 777
- 引证文献数: 0