Blind Detection of Tampered JPEG Image Based on Blocking Artifacts of Double Quantization

Abstract:

In this paper, a novel approach is proposed to detect JPEG image forgery based on the inconsistencies of quantization tables. The primary quantization table of tempered image is firstly estimated. Then the blocking artifact measure (BAM) of each 8×8 DCT coefficient block is computed using the estimated quantization matrix, and the BAMs of all blocks are sorted in descending order. Select a certain number of blocks of which the BAM is in the front of the order. Labeling the connected components of blocks in order to eliminate some abnormal blocks and gain the connected doctored blocks. Experimental results demonstrate the validity of the proposed approach.

• Periodical: Applied Mechanics and Materials (Volumes 263-266)
• Pages: 3104-3107
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.263-266.3104
• Online since: December 2012
• Authors: Shao Bao Wu, Dong Sheng Shen
• Keywords: Blocking Artifact Measure, Connected Components, Double Quantization, Image Tampering, Primary Quantization Matrix

References

[1] H.Farid: IEEE Signal Processing Magazine, Vol. 26(2009), p.16.

[2] A.C. Popescu, H. Farid, in: Proceedings of the 6th International Workshop on Information Hiding, Canada, Springer (2004).

[3] H.Farid: IEEE Transactions on Information Forensics and Security, Vol. 4(2009), p.154.

[4] Luo W Q, Qu Z H, Huang J W, et al. in: Proceedings of ICASSP. New York: IEEE Computer Society (2007).

[5] He Junfeng, Lin Zhouchen, Wang Lifeng, et al, in: Lecture Notes in Computer Science Graze, Australia: Springer (2006).

[6] Ye Shuiming, Sun Qibin, in: Proceedings of ICME. New York: IEEE Computer Society (2007).

[7] J. Lukas, J. Fridrich, in: Proceedings of Digital Forensic Research Workshop (2003).

[8] Information on http://www.philsallee.com/jpegtbx/ index.html.