Tampering Detection for Composite Images Based on Re-sampling and JPEG Compression

被引：0

作者：

Zuo, Juxian ^{[1
]}

Pan, Shengjun ^{[1
]}

Liu, Benyong ^{[1
]}

Liao, Xiang ^{[2
]}

机构：

[1] Guizhou Univ, Inst Intelligent Informat Proc, Sch Comp Sci & Informat, Guiyang 550025, Peoples R China

[2] Minist Publ Secur PRC, Key Lab Sound Image Mat Examinat & Authenticat, Guiyang 550001, Peoples R China

来源：

2011 FIRST ASIAN CONFERENCE ON PATTERN RECOGNITION (ACPR) | 2011年

基金：

国家教育部博士点专项基金资助;

关键词：

composite image; re-sampling; JPEG compression; tampering detection; EXPOSING DIGITAL FORGERIES;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Image forgery detection is an important issue in digital forensics. We propose in this manuscript an algorithm for image composite tampering detection. In the algorithm, we first divide a possibly composite image into overlapped blocks, then define and extract a block measure factor which contains both re-sampling characteristics and JPEG compression characteristics for each block, and apply the block measure factor to discriminate the tampered regions from non-tampered regions. Experimental results show that with the proposed algorithm, composite images are correctly recognized, and at the same time, the tampered regions are determined effectively.

引用

页码：169 / 173

页数：5

共 9 条

[1] Identification of bitmap compression history: JPEG detection and quantizer estimation [J].