针对现有的由渲染软件生成的计算机生成(CG)图像和由数码相机生成的自然图像(PIM)的鉴别方法鉴定准确率不高的问题,提出一种鉴别准确率更高的改进方法。通过离散小波变换(DWT)、直方图计算、离散傅里叶变换(DFT)、统计特征提取并结合正弦滤波特征,提高CG图像与PIM图像的鉴别准确率,为图像取证、数字防伪鉴别等提供可靠的依据。实验表明,该算法具有高度的准确性,鉴别准确率达到95.36%,满足工程应用的需求。
An improved method with a higher identifying accuracy rate is presented to identify computer graphics(CG) generated by rendering software and photographic images(PIM) acquired by digital cameras.By discrete wavelet transform(DWT),histogram calculating,discrete Fourier transform(DFT),statistical feature extraction and combination with a sine-filtering feature,the identifying accuracy rate about CG and PIM is improved.This provides credible supports for some relative domains such as image forensics and digital anti-fake identification etc.Experiments prove that the algorithm has a high accuracy,and the identifying accuracy rate comes to 95.36%,so it meets the engineering applications.