一种估计JPEG双重压缩原始量化步长的新方法

王俊文; 刘光杰; 戴跃伟; 周琳娜; 郭云彪

doi:10.3724/SP.J.1146.2007.01968

一种估计JPEG双重压缩原始量化步长的新方法

doi: 10.3724/SP.J.1146.2007.01968

基金项目:

江苏省自然科学基金(BK2008403)，中国博士后基金(20070421017)，中国信息安全产品评测认证中心115基金和江苏省青蓝工程中青年学术带头人培养基金资助课题

计量
- 文章访问数: 3792
- HTML全文浏览量: 87
- PDF下载量: 1171
- 被引次数: 0
出版历程
- 收稿日期: 2007-12-25
- 修回日期: 2008-05-12
- 刊出日期: 2009-04-19

A New Method for Estimating the Primary Quantization Step of JPEG Double-Compression

摘要

摘要: 该文提出了一种双重压缩后JPEG图像的原始量化步长的估计方法。该方法根据两次量化步长之间的大小关系分3种情况进行讨论。当原始量化步长大于第2次量化步长，提出了直接利用直方图计算的新方法；为解决原始量化步长是第2次量化步长因子和傅里叶频谱分析中的多值问题，提出了采用0.98缩放来近似未压缩图像的方法。本文方法能给出第二次量化步长为一次量化步长倍数时的估计，并利用频谱分析的结果降低了计算的复杂度，实验结果表明本文方法有较高的估计准确度。
- 图像处理;双重量化;直方图;重构;匹配
Abstract: This paper presents a method for estimating the primary quantization step of double compressed JPEG image. The method in accordance with the relationship between the two quantization steps contains three situations. When the primary quantization step is larger than the second quantization step, the histogram can be used to calculate directly. To resolve the issue of multi-value caused by Fourier spectrum analysis and the problem that the primary quantization step is the factor of the second quantization step, the uncompressed image can be approximated by narrowing the image by 98%. This method can resolve the problem that the primary quantization step is the factor of the second quantization step, and uses spectrum analysis to reduce the complexity. The experimental results show that this method has higher accuracy of the estimation.

HTML全文

参考文献(1)

Sencar H T and Memon Nasir. Overview of state-of-art indigital image forensics. Part of Indian Statistical InstitutePlatinum Jubilee Monograph series titled 'Statistical Scienceand Interdisciplinary Research,' World Scientific Press, 2008.[2]Luo Weiqi, Qu Zhenhua, Pan Feng, and Huang Jiwu. Asurvey of passive technology for digital image forensics[J].Front.Comute. Sci. China.2007, 1(2):166-179[3]Fan Z and de Queiroz R L. Identification of bitmapcompression history: JPEG detection and quantizerestimation[J].IEEE Trans. on Image Processing.2003, 12(2):230-235[4]Fridrich J, Goljan M, and Du R. Steganalysis based on JPEGcompatibility. Special session on Theoretical and PracticalIssues in Digital Watermarking and Data Hiding, MultimediaSystems and Applications IV, Denver, Co, USA, August19-24, 2001: 275-280.[5]Fridrich J, Goljan M, and Hogea D. Attacking the outguess.Proc. Multimedia and Security, Workshop at ACMMultimedia, Juan-les-Pins, France, December 6, 2002: 3-6.[6]Fridrich J, Goljan M, and Hogea D. Steganalysis of JPEGimages: Breaking the F5 algorithm. Proc. 5th InternationalWorkshop on Information Hiding, Noordwijkerhout, TheNederlands, October 7-9, 2002: 310-323.[7]Lukas J and Fridrich J. Estimation of primary quantizationmatrix in double compressed JPEG images. Proc. Of DFRWS,Cleveland, OH, USA, August 6-8, 2003: 67-84.[8]Popescu A and Farid H. Statistical tools for digital forensics.6th International Workshop on Information Hiding, Toronto,Canada, May, 2004: 128-147.[9]He J, Lin Z, Wang L, and Tang X. Detecting doctored JPEGimages via DCT coefficient analysis. Proc. of ECCV, Berlin,2006: 423-435.

施引文献

资源附件(0)

访问统计