针对一类JPEG图像伪造的被动盲取证

图像合成是一种最常见的图像伪造手段。该文针对一类JPEG图像合成伪造，根据篡改区域与非篡改区域块效应的不一致性，提出了一种简单有效的检测算法。首先利用估计的一次压缩质量因子对待检测图像进行裁剪再压缩，然后通过计算压缩前后的失真提取图像的块效应指数映射图，最后采用图像分割的方法实现篡改区域的自动检测与定位。实验结果表明，算法对于各种质量的JPEG图像和较小的篡改区域均能有效检测，当二次压缩与一次压缩的质量因子之差在15以上，虚警率控制在5%以内时，检测率可达93%以上。

Passive-Blind Forensics for a Class of JPEG Image Forgery

Compositing is the most common form of digital image forging. For a class of JPEG image forgery, a simple and effective detecting algorithm is proposed on the basis of inconsistencies of blocking artifacts between tampered region and non-tampered region. Firstly, a testing image is cropped and recompressed with the estimated primary quality factor. Secondly, blocking artifacts factor map is extracted via computing the distortion caused by recompression. Finally, automatic detection and location of tampered region is realized by using image segmentation. Experimental results demonstrate that the proposed approach is applicable to images of various qualities and lesser-tampered regions. When the difference of quality factors between double compression and single compression is greater than 15, the accuracy reaches above 93% with a less than 5% false positive rate.