针对目前复制-粘贴篡改盲检测算法对光照变换操作鲁棒性较差的问题,提出了一种基于混合灰度序模式(Mixed Intensity Order Pattern,MIOP)的复制-粘贴篡改盲鉴别算法。首先,对待检测图像提取高斯差分区域(Difference of Gaussians,DOG)...针对目前复制-粘贴篡改盲检测算法对光照变换操作鲁棒性较差的问题,提出了一种基于混合灰度序模式(Mixed Intensity Order Pattern,MIOP)的复制-粘贴篡改盲鉴别算法。首先,对待检测图像提取高斯差分区域(Difference of Gaussians,DOG)。其次,利用MIOP特征描述区域。最后,匹配特征并利用RANSAC(RANdom SAmple Consensus)去除误匹配,确定图像的复制-粘贴篡改区域。实验结果表明:本文算法不仅对几何变换和光照操作的检测率较高,且对高斯模糊、噪声和JPEG重压缩等后处理操作鲁棒性较好。展开更多
Most existing methods for image copy-move forgery detection(CMFD)operate on grayscale images. Although the keypoint-based methods have the advantages of strong robustness and low computational cost,they cannot identif...Most existing methods for image copy-move forgery detection(CMFD)operate on grayscale images. Although the keypoint-based methods have the advantages of strong robustness and low computational cost,they cannot identify the flat duplicated regions without reliable extracted features. In this paper, we propose a new CMFD method by using speeded-up robust feature(SURF)in the opponent color space. Our method starts by converting the inspected image from RGB to the opponent color space. The color gradient per pixel is calculated and taken as the work space for SURF to extract the keypoints. The matched keypoints are clustered and their geometric transformations are estimated. Finally, the false matches are removed. Experimental results show that the proposed technique can effectively expose the duplicated regions with various transformations, even when the duplication regions are flat.展开更多
文摘针对目前复制-粘贴篡改盲检测算法对光照变换操作鲁棒性较差的问题,提出了一种基于混合灰度序模式(Mixed Intensity Order Pattern,MIOP)的复制-粘贴篡改盲鉴别算法。首先,对待检测图像提取高斯差分区域(Difference of Gaussians,DOG)。其次,利用MIOP特征描述区域。最后,匹配特征并利用RANSAC(RANdom SAmple Consensus)去除误匹配,确定图像的复制-粘贴篡改区域。实验结果表明:本文算法不仅对几何变换和光照操作的检测率较高,且对高斯模糊、噪声和JPEG重压缩等后处理操作鲁棒性较好。
基金Supported by the Natural Science Foundation of Tianjin(No.15JCYBJC15500)
文摘Most existing methods for image copy-move forgery detection(CMFD)operate on grayscale images. Although the keypoint-based methods have the advantages of strong robustness and low computational cost,they cannot identify the flat duplicated regions without reliable extracted features. In this paper, we propose a new CMFD method by using speeded-up robust feature(SURF)in the opponent color space. Our method starts by converting the inspected image from RGB to the opponent color space. The color gradient per pixel is calculated and taken as the work space for SURF to extract the keypoints. The matched keypoints are clustered and their geometric transformations are estimated. Finally, the false matches are removed. Experimental results show that the proposed technique can effectively expose the duplicated regions with various transformations, even when the duplication regions are flat.
