基于CFA痕迹的图像局部篡改取证

Image Local-Forgery Detection based on CFA Artifacts

提出了一种能够对数码相机成像的图像进行取证的新方法,可以精确定位到图像中伪造区域的位置。这种处理中,假定图像成像前经过颜色滤波矩阵(CFA),而图像的篡改操作会破坏CFA痕迹,这种假定对大多数商业相机所生成的图像是成立的。因此,在此基础上,找到了一种新的特征,可判定CFA痕迹在局部区域是否存在。这种特征在不需要预测该图像可能被篡改区域的情况下,可以计算很小的图像块(2×2)的伪造概率。实验部分对不同的相机品牌、不同的去马赛克算法进行验证,结果证明该方法检测率高且计算复杂度低。

A forensic tool able to discriminate between original and forged regions in an image captured by a digital camera is proposed. In this process, the image is assumed to be acquired by using CFA（Color Filter Array）, and that the tampering would remove the artifacts due to the demosaicing algorithm. This assumption is true for images generated by most commercial cameras. Based on this, a new feature is found and this feature can be used to determine whether CFA traces exist in the local region, allows the calculation of spurious probabilities of very small image blocks（2 × 2） without predicting the potentially-tampered regions of the image. Experiments on different camera brands and different mosaic algorithms indicate that the proposed method has fairly high detection-rate and low computational complexity.