基于图像表面梯度的翻拍检测

经过翻拍后的篡改图像能够轻易绕过现有的图像篡改检测系统,这对图像真实性的检测构成严重的安全威胁.本文基于翻拍过程中两次使用相机进行拍摄而引入的非线性响应,使用Lib-SVM设计图像分类器判别翻拍图像和真实图像.翻拍后图像表面梯度值与真实图像相比会产生非线性变化,这使翻拍图像表面梯度值产生异常.本文基于图像表面梯度特性提取相关特征值,使用支持向量机分类器进行翻拍图像和真实图像的判别.实验结果表明,本文提出的特征分类效果良好,可以正确检测翻拍图像.

Recaptured image detection based on surface gradient

For most current image forensic systems,the tamper evidence would be invalid in recaptured image,so the authenticity of the image detected is posed a security threat.This paper proposed a separation method between natural images and recaptured photo images,based on surface gradient of image.In the process of image recapturing,because of using camera twice,recaptured image would have a different non-linear response compared to true image.To prevent the image recapturing attack,we propose a set of statistical features,which capture the common anomalies introduced in the camera recapturing process on surface gradient.With a support vector machine classifier,experiment results show that features work proposed can correctly detect recaptured photo.