The identification of real image and photorealistic computer graphics is one of the significant problems in blind digital image forensics. A new method to distinguish the real image from photorealistic computer graphics is presented. This method is based on the differences of high frequency and noise between the two images. The LIBSVM is introduced to construct a classifier to do the differentiating job. The experiment results show that the method can discriminate the two types of images effectively.%真实图像和计算机图像的鉴别是数字图像盲取证研究的重要问题之一.分析、研究了真实图像和计算机图像的成像原理,并发展了对这两类图像鉴别的数字方法.该方法从两类图像的高频信息及噪声形成的区别进行特征提取,并利用LIBSVM进行分类识别,实验表明此方法在鉴别图像类型中可以取得较好的分辨效果.
展开▼
