Transcript F119 Field Margin Increase Study

Image Tampering Detection Using
Bayesian Analytical Methods
04/11/2005
As presented by Jason Kneier
ELEN E6886
Spring 2005
The Problem
• Common image processing tools are capable of
creating forgeries undetectable to the eye
• Data can also be hidden in regions of an image where
it is less likely to perturb the original image
The Solution
• Develop a statistical method to detect tampering and forgeries of
images
Proposal
• Use a Bayesian framework to determine authenticity of images
based on computed feature vectors of image statistics
Methods
Feature vectors of interest:
• Wavelet decomposition
• Biocoherence
System Diagram
Input image
Wavelet Decomposition
into feature vectors
Region is authentic
Bayesian analysis of
feature vectors
Region has been tampered with
Outputs
Determine locations of suspected tampering, and degree of
confidence in determination
References
[1] A. C. Popescu and H. Farid, “Exposing Digital Forgeries by
Detecting Traces of Re-sampling, “ IEEE Transactions on Signal Processing,
53(2):758-767, 2005.
[2] A.C. Popescu and H. Farid, “Statistical Tools for Digital Forensics,” 6th
International Workshop on Information Hiding, Toronto, Canada, 2004.
[3] S. Lyu and H. Farid, “How Realistic is Photorealistic?,” IEEE Transactions on
Signal Processing, 53(2):845-850, 2005.
[4] Tian-Tsong Ng, Shih-Fu Chang, “Blind Detection of Photomontage using Higher
Order Statisics,” Online:
http://www.ee.columbia.edu/~qibin/papers/qibin2004_iscas_1.pdf, Columbia
University, 2004.
[5] R. Duda, P. Hart and D. Stork, Pattern Classification. New York, John Wiley &
Sons, 2001.
[6] T. Cover and J. Thomas, Elements of Information Theory. New York, John Wiley &
Sons, 1991.
[7] W. Pratt, Digital Image Processing. New York, John Wiley & Sons, 2001.
[8] A. Papoulis and S. Pillai, Probability, Random Variables and Stochastic Processes.
Boston, McGraw Hill, 2002.