局部特征在数字图像同源同源篡改取证中的应用研究摘要随着计算机和网络技术的不断发展,数字图像资源的使用率居高不下。在图像媒体得到广泛应用的同时,应各类图像使用的要求,图像处理技术飞速发展,出现了很多图像编辑软件,同源篡改图像变得相当容易,导致图像的安全性受到严重挑战。因此,用于图像信息安全的相关认证技术应用而生,并且迅速发展。克隆操作和拼接操作可以看做在同一副图像或不同图像中生成了原始图像中部分图像的拷贝,可以考虑用局部特征拷贝检测的思想解决这两张操作的检测问题。因此,本文提出了基于局部特征拷贝检测的盲取证算法。一方面,通过特征融合提取精准特征,保证检测的准确性;另一方面,通过基于机器学习的感知哈希将特征映射为紧凑的哈希码,提高检测的效率。但展望海量数字图像在互联网高速传播的未来,仍需继续完善数字图像盲取证技术框架、探索数字图像真实性相关的图像语义、研究利用互联网用户反馈信息帮助数字图像取证的方法等来提高面向真实性鉴别的数字图像盲取证技术的性能。关键字:局部特征;数字图像;同源同源篡改;取证ABSTRACTWiththecontinuousdevelopmentofcomputerandnetworktechnology,theuserateofdigitalimageresourcesishigh.Withthewideapplicationofimagemedia,variousimagerequestsandimageprocessingtechnologieshavebeendevelopingrapidly.Manyimageeditingsoftwareshaveappeared.Theimageofhomologoustamperinghasbecomequiteeasy,whichleadstoseriouschallengestoimagesecurity.Therefore,therelevantauthenticationtechnologyforimageinformationsecurityhasbeendevelopedanddevelopedrapidly.Cloningoperationandstitchingoperationcanberegardedascopiesoforiginalimagesinthesameimageordifferentimages.Wecanconsidertheideaoflocalfeaturecopydetectiontosolvethedetectionproblemsofthesetwooperations.Therefore,ablindforensicsalgorithmbasedonlocalfeaturecopydetectionisproposedinthispaper.Ontheonehand,accuratefeaturesareextractedthroughfeaturefusiontoensuretheaccuracyofdetection.Ontheotherhand,theperceptionofHashibasedonmachinelearningmapsthefeaturetocompactHashicode,whichimprovestheefficiencyofdetection.ButtheprospectofmassivedigitalimageontheInternetpropagationspeedofthefuture,stillneedtocontinuetoimprovethedigitalimageforensicstechnologyframework,exploretheauthenticityofdigitalimage,researchrelatedtotheuseoftheInternetuserperformancefeedbackmethodofdigitalimageforensicshelptoimprovetheblindforensicsofdigitalimageauthenticitydetection.Keywords:localfeatures;digitalimages;homologousandhomologoustampering;Forensics目录1、绪论...................................................41.1研究的背景与意义...................................41.2本文的主体架构.....................................52、盲取证基本理论和关键技术...............................52.1数字图像取证分类...................................52.2数字图像主动取证技术...............................62.2.1数字水印......................................62.2.2数字签名......................................72.3数字图像被动取证(盲取证)技术.....................73、基于局部特征拷贝检测的盲取证研究.......................83.1基于局部特征拷贝检测的被动取证框架.................83.2局部特征图像拷贝检测算法...........................83.3基于特征融合的特征提取算法.........................93.4基于机器学习的感知哈希算法........................123.5实验结果与分析....................................173.5.1局部特征拷贝检测实验与结果分析...............173.5.2被动取证实验与结果分析.......................203.5.3取证算法性能分析.............................22结语.....................................................