基于多残差马尔科夫模型的图像拼接检测

摘要

针对传统马尔科夫特征计算差值矩阵的方式单一、拼接检测鲁棒性不强的问题,提出彩色多残差马尔科夫特征拼接检测模型.该模型引入隐写检测模型(Rich Models for Steganalysis,SRM)中的多种残差类型来改进传统马尔科夫特征,从R,G,B3个通道分别提取10种不同类型的马尔科夫特征,训练30个独立的SVM分类器,最后通过决策判断进行分类预测.该方法在哥伦比亚大学彩色拼接检测库上达到了95.40％的准确率.%Aiming at the problem that calculating the difference matrix singly for the traditional Markov feature is not robust to the splicing detection,this paper presented a color multi-residual type Markov feature for splicing detection.This method introduces the residual model from rich models for steganalysis(SRM) to improve the traditional Markov features,respectively extracts 10 different types of Markov features from three color channels,and trains 30 unique SVM classifiers to make the classification through the proposed decision-making algorithm.This method achieves the accuracy of 95.40％ at Columbia image splicing detection evaluation dataset.