一种新的自然图像和计算机图形鉴别方法  

A NOVEL APPROACH FOR DISTINGUISHING PHOTO IMAGES AND COMPUTER GRAPHICS

作　　者：张震[1,2] 康吉全[2] 平西建[1] 任远[2] 

机构地区：[1]信息工程大学信息工程学院,河南郑州450002 [2]郑州大学电气工程学院,河南郑州450001

出　　处：《计算机应用与软件》2009年第12期30-33,共4页Computer Applications and Software

基　　金：国家自然基金项目(60473022)

摘　　要：提出一种自然图像和计算机生成图形的鉴别方法。借用模式识别中的二分类概念,采用块离散余弦变换和统计矩特征量来建立模型,以捕获自然图像和计算机生成图形的统计差异,选用支持向量机作为分类器进行训练和测试。实验结果表明,该方法具有精确度高、应用面广的优点,在自然图像和计算机生成图形的鉴别中有着广阔的前景。This paper proposes a new detection scheme to discriminate computer graphics from photo images. The detection can be regarded as a two-class pattern recognition problem and the model is established based on block discrete cosine transform （DCT） and statistical moment eigenvalue extracted from the given test image. This model can capture statistical differences between photo images and computer graphics. Kernel-based Support Vector Machine （ SVM ） is chosen as a classifier to train and test the given images. Experimental results demonstrate that this new detection scheme has some advantages of high-accuracy and widely-application, indicating that the proposed approach possesses promising capability in discrimination of computer graphics from photo images.

关 键 词：自然图像 计算机生成图形 统计矩特征 支持向量机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP181[自动化与计算机技术—计算机科学与技术]