Research article Special Issues

Joint statistics matching for camera model identification of recompressed images

  • Received: 24 January 2019 Accepted: 09 May 2019 Published: 31 May 2019
  • Source camera identification has been well studied in laboratory environment where the training and test samples are all original images without recompression. However, image compression is quite common in the real world, when the training and test images are double JPEG compressed with different quantization tables, the identification accuracy of existing methods decreases dramati- cally. To address this challenge, we propose a novel iterative algorithm namely joint first and second order statistics matching (JSM) to learn a feature projection that projects the training and test fea- tures into a low dimensional subspace to reduce the shift caused by image recompression. Inspired by transfer learning, JSM aims to learn a new feature representation from original feature space by simultaneously matching the first and second order statistics between training and test features in a principled dimensionality reduction procedure. After the feature projection, the divergence between training and test features caused by recompression is reduced while the discriminative properties are preserved. Extensive experiments on public Dresden Image Database verify that JSM significantly outperforms several state-of-the-art methods on camera model identification of recompressed images.

    Citation: Bo Wang, Yabin Li, Xue Sui, Ming Li, Yanqing Guo. Joint statistics matching for camera model identification of recompressed images[J]. Mathematical Biosciences and Engineering, 2019, 16(5): 5041-5061. doi: 10.3934/mbe.2019254

    Related Papers:

  • Source camera identification has been well studied in laboratory environment where the training and test samples are all original images without recompression. However, image compression is quite common in the real world, when the training and test images are double JPEG compressed with different quantization tables, the identification accuracy of existing methods decreases dramati- cally. To address this challenge, we propose a novel iterative algorithm namely joint first and second order statistics matching (JSM) to learn a feature projection that projects the training and test fea- tures into a low dimensional subspace to reduce the shift caused by image recompression. Inspired by transfer learning, JSM aims to learn a new feature representation from original feature space by simultaneously matching the first and second order statistics between training and test features in a principled dimensionality reduction procedure. After the feature projection, the divergence between training and test features caused by recompression is reduced while the discriminative properties are preserved. Extensive experiments on public Dresden Image Database verify that JSM significantly outperforms several state-of-the-art methods on camera model identification of recompressed images.


    加载中


    [1] M. C. Stamm, M. Wu and K. J. R. Liu, Information forensics: An overview of the first decade, IEEE Access, 1 (2013), 167–200.
    [2] A. Piva, An overview on image forensics, Isrn Signal Process., 2013.
    [3] M. Kirchner and T. Gloe, Forensic camera model identification, in Handbook of Digital Forensics of Multimedia Data and Devices, Chichester, U.K.: Wiley, 2015, 329–374.
    [4] J. Lukáš, J. Fridrich and M. Goljan, Determining digital image origin using sensor imperfections, in Proc. SPIE, San Jose, CA, USA, 2005, 249–260.
    [5] M. Chen, J. J. Fridrich and M. Goljan, Digital imaging sensor identification (further study)., in Proc. SPIE, San Jose, CA, USA, 2007, 65050P–65050P13.
    [6] C. T. Li, Source camera identification using enhanced sensor pattern noise, IEEE Trans. Inf. Foren. Secur., 5 (2010), 280–287.
    [7] X. Kang, Y. Li, Z. Qu, et al., Enhanced source camera identification performance with a camera reference phase sensor, IEEE Trans. Inf. Foren. Secur., 7 (2012), 393–402.
    [8] A. Lawgaly, F. Khelifi and A. Bouridane, Weighted averaging-based sensor pattern noise esti-mation for source camera identification, in Proc. IEEE Int. Conf. Image Process., Paris, France, 2014, 5357–5361.
    [9] A. Lawgaly and F. Khelifi, Sensor pattern noise estimation based on improved locally adaptive dct filtering and weighted averaging for source camera identification and verification, IEEE Trans. Inf. Foren. Secur., 12 (2017), 392–404.
    [10] Y. Hu, C. T. Li and Z. Lai, Fast source camera identification using matching signs between query and reference fingerprints, Mult. Tool. Appl., 74 (2015), 7405–7428.
    [11] S. Bayram, H. T. Sencar and N. Memon, Sensor fingerprint identification through composite fingerprints and group testing, IEEE Trans. Inf. Foren. Secur., 10 (2015), 597–612.
    [12] D. Valsesia, G. Coluccia, T. Bianchi, et al., Compressed fingerprint matching and camera identi-fication via random projections, IEEE Trans. Inf. Foren. Secur., 10 (2015), 1472–1485.
    [13] R. Li, C. T. Li and Y. Guan, Inference of a compact representation of sensor fingerprint for source camera identification, Patt. Recogn., 74 (2018), 556 – 567.
    [14] Y. Hu, C. T. Li, C. Zhou, et al., Source camera identification issues: forensic features selection and robustness, Int. J. Digit. Crime Foren., 3 (2011), 1–15.
    [15] M. Kharrazi, H. T. Sencar and N. Memon, Blind source camera identification, in Proc. IEEE Int. Conf. Image Process., vol. 1, Singapore, 2004, 709–712.
    [16] K. San Choi, E. Y. Lam and K. K. Wong, Source camera identification by jpeg compression statistics for image forensics, in Proc. IEEE Region 10 Conf., Hong Kong, China, 2006, 1–4.
    [17] B. Wang, Y. Guo, X. Kong, et al., Source camera identification forensics based on wavelet features, in Proc. IEEE Fifth Int. Conf. Intelligent Inf. Hiding and Multimedia Signal Process., Kyoto, Japan, 2009, 702–705.
    [18] S. Bayram, H. Sencar, N. Memon, et al., Source camera identification based on cfa interpolation, in Proc. IEEE Int. Conf. Image Process., vol. 3, Genova, Italy, 2005, 69–72.
    [19] J. S. Ho, O. C. Au, J. Zhou, et al., Inter-channel demosaicking traces for digital image forensics, in Proc. IEEE Int. Conf. Multimedia Expo, Suntec, Singapore, 2010, 1475–1480.
    [20] Y. Hu, C. T. Li, X. Lin, et al., An improved algorithm for camera model identification using inter-channel demosaicking traces, in 8th Int. Conf. Intelligent Inform. Hiding and Multimedia Signal Process., Piraeus, Greece, 2012, 325–330.
    [21] G. Xu and Y. Q. Shi, Camera model identification using local binary patterns, in Proc. IEEE Int. Conf. Multimedia Expo, Melbourne, VIC, Australia, 2012, 392–397.
    [22] B. Xu, X. Wang, X. Zhou, et al., Source camera identification from image texture features, Neurocomputing, 207 (2016), 131–140.
    [23] C. Chen and M. C. Stamm, Camera model identification framework using an ensemble of de-mosaicing features, in Proc. IEEE International Workshop on Information Forensics and Security (WIFS), 2015, 1–6.
    [24] A. Tuama, F. Comby and M. Chaumont, Camera model identification based machine learning approach with high order statistics features, in Proc. 24th European Signal Processing Conference (EUSIPCO), 2016, 1183–1187.
    [25] A. Roy, R. S. Chakraborty, U. Sameer, et al., Camera source identification using discrete cosine transform residue features and ensemble classifier, in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2017.
    [26] L. Zeng, X. Kong, M. Li, et al., Jpeg quantization table mismatched steganalysis via robust dis-criminative feature transformation., in Proc. SPIE Media Watermarking, Security, and Forensics, San Francisco, CA, USA, 2015, 94090U–94090U9.
    [27] S. J. Pan, I. W. Tsang, J. T. Kwok, et al., Domain adaptation via transfer component analysis, IEEE Trans. Neural Netw., 22 (2011), 199–210.
    [28] M. Long, J. Wang, G. Ding, et al., Transfer feature learning with joint distribution adaptation, in Proc. IEEE Int. Conf. Comput. Vision, San Francisco, CA, USA, 2013, 2200–2207.
    [29] X. Li, X. Kong, B. Wang, et al., Generalized transfer component analysis for mismatched jpeg steganalysis, in Proc. IEEE Int. Conf. Image Process., Melbourne, VIC, Australia, 2013, 4432–4436.
    [30] L. Luo, X. Wang, S. Hu, et al., Close yet distinctive domain adaptation, arXiv:1704.04235.
    [31] G. Zhang, B. Wang and Y. Li, Cross-class and inter-class alignment based camera source identi-fication for re-compression images, in Proc. IEEE Int. Conf. Image Graphics, Shanghai, China, 2017.
    [32] A. Gretton, K. M. Borgwardt, M. Rasch, et al., A kernel method for the two-sample-problem, in Proc. Advances Neural Inf. Process. Syst., Vancouver, BC, Canada, 2007, 513–520.
    [33] B. Quanz, J. Huan and M. Mishra, Knowledge transfer with low-quality data: A feature extraction issue, IEEE Trans. Knowl. Data Eng., 24 (2012), 1789–1802.
    [34] E. Zhong, W. Fan, J. Peng, et al., Cross domain distribution adaptation via kernel mapping, in Proc. 15th ACM SIGKDD Int. Conf. on Knowl. Discovery Data Mining, Paris, France, 2009, 1027–1036.
    [35] T. Gloe and R. Böhme, The dresden image database for benchmarking digital image forensics, J. Digit. Forensic Pract., 3 (2010), 150–159.
  • Reader Comments
  • © 2019 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(3040) PDF downloads(473) Cited by(3)

Article outline

Figures and Tables

Figures(4)  /  Tables(5)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog