Random subspace method for source camera identification

Ruizhe Li; Constantine Kotropoulos; Chang Tsun Li; Yu Guan

doi:10.1109/MLSP.2015.7324339

Random subspace method for source camera identification

Ruizhe Li, Constantine Kotropoulos, Chang Tsun Li, Yu Guan

Computing and Mathematics

Research output: Book chapter/Published conference paper › Conference paper

4 Citations (Scopus)

Abstract

Sensor pattern noise is an inherent fingerprint of imaging devices, which has been widely used for source camera identification, image classification, and forgery detection. In a previous work, we proposed a feature extraction method based on the principal component analysis denoising concept, which can enhance the performance of conventional SPN extraction methods. However, this method is vulnerable, because the training samples are seriously affected by the image content. Accordingly, it is difficult to train a reliable feature extractor by using such a training set. To address this problem, a camera identification framework based on the random subspace method and majority voting is proposed in this work. The experimental results show that the proposed solution can suppress the interference from scene details and enhance the performance in terms of the receiver operating characteristic curve.

Original language	English
Title of host publication	2015 IEEE International Workshop on Machine Learning for Signal Processing
Subtitle of host publication	Proceedings of MLSP 2015
Editors	Deniz Erdoğmuş , Murat Akçakaya , Serdar Kozat, Jan Larsen
Publisher	IEEE Computer Society
Pages	1-5
Number of pages	5
Volume	2015-November
ISBN (Electronic)	9781467374545
DOIs	https://doi.org/10.1109/MLSP.2015.7324339
Publication status	Published - 10 Nov 2015
Event	25th IEEE International Workshop on Machine Learning for Signal Processing: MLSP 2015 - Northeastern University, Boston, United States Duration: 17 Sep 2015 → 20 Sep 2015 http://mlsp2015.conwiz.dk/home.htm (Conference website)

Conference

Conference	25th IEEE International Workshop on Machine Learning for Signal Processing
Country	United States
City	Boston
Period	17/09/15 → 20/09/15
Internet address	http://mlsp2015.conwiz.dk/home.htm (Conference website)

Access to Document

10.1109/MLSP.2015.7324339

Fingerprint Dive into the research topics of 'Random subspace method for source camera identification'. Together they form a unique fingerprint.

Cite this

Li, R., Kotropoulos, C., Li, C. T., & Guan, Y. (2015). Random subspace method for source camera identification. In D. Erdoğmuş , M. Akçakaya , S. Kozat, & J. Larsen (Eds.), 2015 IEEE International Workshop on Machine Learning for Signal Processing: Proceedings of MLSP 2015 (Vol. 2015-November, pp. 1-5). [7324339] IEEE Computer Society. https://doi.org/10.1109/MLSP.2015.7324339

@inproceedings{0432714c846f43bfa81293e1ba8491d8,

title = "Random subspace method for source camera identification",

abstract = "Sensor pattern noise is an inherent fingerprint of imaging devices, which has been widely used for source camera identification, image classification, and forgery detection. In a previous work, we proposed a feature extraction method based on the principal component analysis denoising concept, which can enhance the performance of conventional SPN extraction methods. However, this method is vulnerable, because the training samples are seriously affected by the image content. Accordingly, it is difficult to train a reliable feature extractor by using such a training set. To address this problem, a camera identification framework based on the random subspace method and majority voting is proposed in this work. The experimental results show that the proposed solution can suppress the interference from scene details and enhance the performance in terms of the receiver operating characteristic curve.",

keywords = "Digital forensics, PCA denoising, Random subspace method, Sensor pattern noise",

author = "Ruizhe Li and Constantine Kotropoulos and Li, {Chang Tsun} and Yu Guan",

year = "2015",

month = nov,

day = "10",

doi = "10.1109/MLSP.2015.7324339",

language = "English",

volume = "2015-November",

pages = "1--5",

editor = "{Erdoğmu{\c s} }, Deniz and {Ak{\c c}akaya }, Murat and Serdar Kozat and Jan Larsen",

booktitle = "2015 IEEE International Workshop on Machine Learning for Signal Processing",

publisher = "IEEE Computer Society",

address = "United States",

note = "25th IEEE International Workshop on Machine Learning for Signal Processing : MLSP 2015 ; Conference date: 17-09-2015 Through 20-09-2015",

url = "http://mlsp2015.conwiz.dk/home.htm",

}

Li, R, Kotropoulos, C, Li, CT & Guan, Y 2015, Random subspace method for source camera identification. in D Erdoğmuş , M Akçakaya , S Kozat & J Larsen (eds), 2015 IEEE International Workshop on Machine Learning for Signal Processing: Proceedings of MLSP 2015. vol. 2015-November, 7324339, IEEE Computer Society, pp. 1-5, 25th IEEE International Workshop on Machine Learning for Signal Processing, Boston, United States, 17/09/15. https://doi.org/10.1109/MLSP.2015.7324339

Random subspace method for source camera identification. / Li, Ruizhe; Kotropoulos, Constantine; Li, Chang Tsun; Guan, Yu.

2015 IEEE International Workshop on Machine Learning for Signal Processing: Proceedings of MLSP 2015. ed. / Deniz Erdoğmuş ; Murat Akçakaya ; Serdar Kozat; Jan Larsen. Vol. 2015-November IEEE Computer Society, 2015. p. 1-5 7324339.

Research output: Book chapter/Published conference paper › Conference paper

TY - GEN

T1 - Random subspace method for source camera identification

AU - Li, Ruizhe

AU - Kotropoulos, Constantine

AU - Li, Chang Tsun

AU - Guan, Yu

PY - 2015/11/10

Y1 - 2015/11/10

N2 - Sensor pattern noise is an inherent fingerprint of imaging devices, which has been widely used for source camera identification, image classification, and forgery detection. In a previous work, we proposed a feature extraction method based on the principal component analysis denoising concept, which can enhance the performance of conventional SPN extraction methods. However, this method is vulnerable, because the training samples are seriously affected by the image content. Accordingly, it is difficult to train a reliable feature extractor by using such a training set. To address this problem, a camera identification framework based on the random subspace method and majority voting is proposed in this work. The experimental results show that the proposed solution can suppress the interference from scene details and enhance the performance in terms of the receiver operating characteristic curve.

AB - Sensor pattern noise is an inherent fingerprint of imaging devices, which has been widely used for source camera identification, image classification, and forgery detection. In a previous work, we proposed a feature extraction method based on the principal component analysis denoising concept, which can enhance the performance of conventional SPN extraction methods. However, this method is vulnerable, because the training samples are seriously affected by the image content. Accordingly, it is difficult to train a reliable feature extractor by using such a training set. To address this problem, a camera identification framework based on the random subspace method and majority voting is proposed in this work. The experimental results show that the proposed solution can suppress the interference from scene details and enhance the performance in terms of the receiver operating characteristic curve.

KW - Digital forensics

KW - PCA denoising

KW - Random subspace method

KW - Sensor pattern noise

UR - http://www.scopus.com/inward/record.url?scp=84960863809&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84960863809&partnerID=8YFLogxK

U2 - 10.1109/MLSP.2015.7324339

DO - 10.1109/MLSP.2015.7324339

M3 - Conference paper

AN - SCOPUS:84960863809

VL - 2015-November

SP - 1

EP - 5

BT - 2015 IEEE International Workshop on Machine Learning for Signal Processing

A2 - Erdoğmuş , Deniz

A2 - Akçakaya , Murat

A2 - Kozat, Serdar

A2 - Larsen, Jan

PB - IEEE Computer Society

T2 - 25th IEEE International Workshop on Machine Learning for Signal Processing

Y2 - 17 September 2015 through 20 September 2015

ER -