Multimedia Security And Forensics Authentication of Digital Images Sarah Summers Sarah Wahl CS525 Semester Project Spring 2006

Motivation Seeing is believing … or is it?

Easy to be deceived

Goals  Identify image tampering methods.  Assess methods available for protecting images.  Assess image authentication techniques.  Identify directions for future work.

Categories of Image Tampering  There are three main categories of image tampering:  Enhancing  Compositing  Copy/Move

Enhancing  Changing the color of objects  Changing the weather conditions  Blurring out objects

Compositing Combining two or more images to create a new image

Copy-Move Copying regions of the original image and pasting into other areas. The yellow area has been copied and moved to conceal the truck.

What can be done to protect digital images?  Watermarking  Fragile watermarks  Semi-fragile watermarks  Self-embedding watermarks  Digital cameras with watermarking capabilities  Digital Fingerprinting/Signatures  Digital cameras with fingerprinting capabilities

Digital Watermarking  The basic concept of digital watermarking an image is that a low level signal is placed directly into the image data.  Any manipulation of the image will impact the watermark and subsequent retrieval of the watermark and examination of its condition will indicate if tampering has occurred.

Fragile Watermarks  Fragile watermarks are designed to detect every possible change in pixel values.  Variety of Techniques but in most cases, the watermark is embedded in the least significant bit (LSB) of the image.  Advantages: Pick up all image manipulations – malicious and non- malicious  Disadvantages: Too sensitive

Semi-Fragile Watermarks  They are robust, to a certain extent, and are less sensitive to pixel modifications.  Techniques:  Divide image into blocks and utilize bits from each block to calculate a spread spectrum noise like signal which is combined with DCT coefficients and inserted as a watermark.  Divide image into blocks, construct watermark in DCT domain from pseudo-random zero-mean unit variance Gaussian numbers, take the inverse DCT and insert into the image.  Advantage: less sensitive than fragile watermarks

Self-Embedding  Tampered images result in lost information. The previous techniques will only detect and localize areas of interest when authentication is carried out.  Self-embedding allows tamper detection and recovery of missing information.  General concept is that the image is embedded in itself in an encrypted form.  Advantage: Potential for original data to be retrieved.  Disadvantage: Tampering with the image can remove blocks of the original image making retrieval of content impossible

Digital Cameras with Watermarking Capabilities  Watermarking based on secret key, block ID and content. The image is divided into blocks and each block watermarked using a frequency based spread spectrum technique incorporating the secret key, block ID and block content.  Image of photographers iris is combined with the camera ID, the hash of the original image and other details specific to the camera.

Digital Fingerprints/Signatures  Based on the concept of public key encryption.  Hashed version of image is encrypted using a private key.  Encrypted file provides a unique signature/fingerprint of the image which can be used to authenticate by decryption with public key.  Mainly used in transmission of images.

Digital Cameras with Fingerprinting Capabilities  Epson Image Authentication System (IAS)  The IAS software in the camera instantly seals the captured images with an invisible digital fingerprint.  Verification of image is achieved by any PC with Image Authentication System software installed

Authentication Techniques  Active Authentication  Rely on the presence of a watermark or fingerprint.  Require knowledge original image  Algorithm/key used to embed the watermark or fingerprint.  Passive Authentication  No requirement of knowledge of original image.  Does not rely of presence of watermark or fingerprint.

Passive Authentication Techniques  Detecting Copy-Move  Detecting Traces of Re-sampling  Detecting Light Inconsistencies

Copy-Move Detection Original ImageTampered Image Exact MatchRobust Match

Copy-Move Detection Original Image Tampered Image PCA Detection

Re-sampling Detection Original ImageTampered Image Periodic pattern in Fourier Transform of altered region Fourier Transform of unaltered region

Inconsistencies in Lighting Genuine Image Tampered Image

Future Research  Development of a better self embedding technique.  Development of an all inclusive passive authentication technique.

Conclusions  Digital image forgeries can be used to deceive the public and the authorities.  They are here to stay.  Until non destructible/ non removal digital watermarks are perfected, passive authentication will remain necessary.  Currently no single passive authentication technique can detect all types of digital forgeries.

References  Hany Farid, Creating and Detecting Doctored and Virtual Images: Implications to The Child Pornography Prevention Act, Technical Report, TR , Dartmouth College, Computer Science.  Detection of Copy-Move Forgery in Digital Images, Jessica Fridrich, David Soukal and Jan Lukas, Proceedings of Digital Forensic Research Workshop, August 2003,  Detection of image alterations using semi-fragile watermarks, E.T. Lin, C. I. Podilchuk, and E.J. Delp,  Semi-fragile watermarking for Telltale Tamper Proofing and Authenticating, H. H. Ko and S. J. Park, Fragile%20Watermarking%20for%20Telltale%20Tamper%20Proofing%20and%20A. doc Fragile%20Watermarking%20for%20Telltale%20Tamper%20Proofing%20and%20A. dochttp:// Fragile%20Watermarking%20for%20Telltale%20Tamper%20Proofing%20and%20A. doc  Methods for Tamper Detection in Digital Images, Jiri Fridrich, Proc. ACM Workshop on Multimedia and Security, Orlando, FL, October 30-31, 1999, pp ,  Information Authentication for a Slippery New Age, S. Walton, Dr. Dobbs Journal, Vol. 20, No. 4, pp 18-26, Apr 1995  Blind Detection of Photomontage using Higher Order Statistics, T. Ng, S. Chang and Q. Sun,

References (continued)  A Digital Watermark, R. van Schyndel, A. Tirkel and C. Osborne, Proceedings of the IEEE International Conference on Image Processing, vol. 2, pp , Austin, Texas, November  A Watermark for Image Integrity and Ownership Verification, P. Wong, IS&T’s 1998 Image Processing, Image Quality, Image Capture, Systems Conference, Portland, Oregon, May 1998, pp. 374 – 379  An Invisible Watermarking Technique for Image Verification, M. Yeung and F. Mintzer, Proc. ICIP’97, Santa Barbara, California 1997  Image watermarking for tamper detection, Jiri Fridrich, Proc. ICIP '98, Chicago, Oct 1998,  Methods for Detecting Changes in Digital Images, J. Fridrich, Proc. of The 6th IEEE International Workshop on Intelligent Signal Processing and Communication Systems (ISPACS'98), Melbourne, Australia, 4-6 November 1998, pp. 173–177,  A Robust Content Based Digital Signature for Image Authentication, M. Schneider and S. Chang, Proceedings of the International Conference on Image Processing, 1996, Volume 3, Issue, Sep 1996 Page(s):

References (continued)  A New Fingerprinting Method for Digital Images, V. Fotopoulos and A. N. Skodras,  Mehdi Kharrazi, Husrev T. Sencar and Nasir Memon, Blind Source Camera Identification, International Conference on Image Processing, 2004, ICIP’04, Volume 1, Oct. 2004, pp  Rotation, Scale and Translation Invariant Digital Image Watermarking, J.J.K. O’Ruanaidh and T. Pun, Proceedings of the ICIP, VOl. 1, pp , Santa Barbara, California, Oct  Secure Digital Camera, Paul Blythe and Jessica Fridrich,  Alin C. Popescu and Hany Farid, Exposing Digital Forgeries in Color Filter Array Interpolated Images, IEEE Transactions on Signal Processing, Vol. 53, Issue 10, Part 2, October 2005, pp  Epson's Image Authentication for digicams,  When is Seeing Believing, W. J. Mitchell, Scientific American, pp , February 1994.

References (continued)  Exposing digital forgeries by detecting inconsistencies in lighting by M. K. Johnson and H. Farid, ACM Multimedia and Security Workshop, New York, NY, 2005,  Exposing Digital Forgeries by Detecting Traces of Re-sampling, A. C. Popescu and H. Farid, IEEE Transactions on Signal Processing, 53(2): , 2005,  Exposing digital forgeries by detecting duplicated image regions, A. C. Popescu and H. Farid, Technical Report , Dartmouth College,  A Tutorial on Principal Components Analaysis, Lindsay I. Smith  Automatic Estimation of the Projected Light Source Direction, P. Nillius and j. –O. Eklundh, Proceddings of the IEEE Computer Science Conference on Computer Vision and Pattern Recognition, 2001  Protection of Digital Images Using Self Embedding, J. Fridrich and M. Goljan, Symposium on Content Security and Data Hiding in Digital Media, New Jersey Institute of Technology, May 14, 1999,  A Model for Image Splicing, T. Ng and S. Chang, ICIP '04. International Conference on Image Processing,. Volume 2, Oct Page(s): Vol.2