计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (1): 179-182.DOI: 10.3778/j.issn.1002-8331.2009.01.055

• 图形、图像、模式识别 • 上一篇    下一篇

不同分类器在图像盲取证中的表现

张 震1,2,边玉琨2,平西建1,张 涛1   

  1. 1.信息工程大学 信息工程学院,郑州 450002
    2.郑州大学 电气工程学院,郑州 450001
  • 收稿日期:2008-06-24 修回日期:2008-08-21 出版日期:2009-01-01 发布日期:2009-01-01
  • 通讯作者: 张 震

Performance of different classifiers in blind image forensics

ZHANG Zhen1,2,BIAN Yu-kun2,PING Xi-jian1,ZHANG Tao1   

  1. 1.Institute of Information Engineering,Information Engineering University,Zhengzhou 450002,China
    2.School of Electrical Engineering,Zhengzhou University,Zhengzhou 450001,China
  • Received:2008-06-24 Revised:2008-08-21 Online:2009-01-01 Published:2009-01-01
  • Contact: ZHANG Zhen

摘要: 为了对数字拼接图像进行盲检测,提出了一种新的拼接图像的检测模型。使用图像质量评价量和统计特征量来建立模型,以得到原始图像和拼接图像之间的统计差异。选用支持向量机和人工神经网络作为分类器分别对该模型进行训练和测试,对拼接图像的盲检测进行了研究。实验结果表明,两种分类器都表现出较高的识别率,该模型在图像拼接检测中有着广阔的前景。

关键词: 数字图像盲取证, 图像拼接检测, 图像质量评价量, 矩特征量, 人工神经网络, 支持向量机

Abstract: To implement image splicing blind detection,this paper proposes a new splicing detection model.The model is built based on statistical moment features and some image quality metrics(IQMs) extracted from the given test image.This model can measure statistical differences between original image and spliced image.Support Vector Machine(SVM) and Artificial Neural Network(ANN) are chosen as the classifiers to train and test the given images based on the proposed model.Experimental results demonstrate that both of the two classifiers have high accuracies and this model possesses promising capability in image splicing detection.

Key words: blind image forensics, image splicing detection, Image Quality Metrics(IQMs), moment features, Artificial Neural Network(ANN), Support Vector Machine(SVM)