图像内容感知缩放的检测方法研究
发布时间:2018-10-04 22:00
【摘要】:针对可用于图像篡改的内容感知缩放技术,提出了一种基于概率Map图统计特征的内容感知缩放检测算法.该算法利用概率Map图以反映图像是否经过内容感知缩放操作,并利用新提出的积分投影与局部统计特征来检测篡改图像.而后利用分类器进行分类训练,从而有效识别基于内容感知缩放操作的图像篡改.实验结果显示,所提算法能够区分出原始图像与篡改图像,并具有较高的正确检测率.
[Abstract]:A content-aware scaling detection algorithm based on the statistical features of probabilistic Map diagrams is proposed for content-aware scaling techniques which can be used for image tampering. The algorithm uses probabilistic Map diagrams to reflect whether the image has undergone content-sensing scaling operations, and uses the newly proposed integral projection and local statistical features to detect tampered images. Then classifier is used for classification training to effectively identify image tampering based on content-aware scaling operation. The experimental results show that the proposed algorithm can distinguish the original image from the tampered image and has a high detection rate.
【作者单位】: 北京交通大学信息科学研究所 现代信息科学与网络技术北京市重点实验室(北京交通大学)
【基金】:国家重点研发计划(2016YFB0800404) 国家自然科学基金(61672090,61332012) 中央高校基本科研业务费专项基金(2015JBZ002)~~
【分类号】:TP391.41
本文编号:2252032
[Abstract]:A content-aware scaling detection algorithm based on the statistical features of probabilistic Map diagrams is proposed for content-aware scaling techniques which can be used for image tampering. The algorithm uses probabilistic Map diagrams to reflect whether the image has undergone content-sensing scaling operations, and uses the newly proposed integral projection and local statistical features to detect tampered images. Then classifier is used for classification training to effectively identify image tampering based on content-aware scaling operation. The experimental results show that the proposed algorithm can distinguish the original image from the tampered image and has a high detection rate.
【作者单位】: 北京交通大学信息科学研究所 现代信息科学与网络技术北京市重点实验室(北京交通大学)
【基金】:国家重点研发计划(2016YFB0800404) 国家自然科学基金(61672090,61332012) 中央高校基本科研业务费专项基金(2015JBZ002)~~
【分类号】:TP391.41
【相似文献】
相关硕士学位论文 前2条
1 万鹏;基于FPGA的视频叠加融合系统设计与实现[D];哈尔滨工程大学;2014年
2 职承伟;复印机扫描控制器的设计与实现[D];西安电子科技大学;2017年
,本文编号:2252032
本文链接:https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2252032.html