detect digital image forgeries

55
Detect Digital Detect Digital Image Forgeries Image Forgeries Ting-Wei Hsu Ting-Wei Hsu

Upload: dior

Post on 11-Jan-2016

59 views

Category:

Documents


3 download

DESCRIPTION

Detect Digital Image Forgeries. Ting-Wei Hsu. History of photo manipulation. 1860 the portrait of Lincoln is a composite of Lincoln ’ s head and John Calhoun ’ s body. History of photo manipulation. 1917: “ Cottingley fairies. History of photo manipulation. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Detect Digital Image Forgeries

Detect Digital Image Detect Digital Image ForgeriesForgeries

Detect Digital Image Detect Digital Image ForgeriesForgeries

Ting-Wei HsuTing-Wei Hsu

Page 2: Detect Digital Image Forgeries

History of photo manipulation

• 1860 the portrait of Lincoln is a composite of Lincoln’s head and John Calhoun’s body

Page 3: Detect Digital Image Forgeries

History of photo manipulation

• 1917: “Cottingley fairies

Page 4: Detect Digital Image Forgeries

History of photo manipulation

• 1930s: Stalin had disgraced comrades airbrushed out of his pictures

Page 5: Detect Digital Image Forgeries

History of photo manipulation

• 1936: same story with Mao

Page 6: Detect Digital Image Forgeries

History of photo manipulation

• 1936: same story with Mao

Page 7: Detect Digital Image Forgeries

History of photo manipulation

• Oprah Winfrey head on Ann-Margret

Page 8: Detect Digital Image Forgeries

History of photo manipulation

• 1994: O.J. Simpson’s mug shot modified to appear more menacing

Page 9: Detect Digital Image Forgeries

History of photo manipulation

Page 10: Detect Digital Image Forgeries

History of photo manipulation

• April 2003: This digital composite of a British soldier in Basra, gesturing to Iraqi civilians urging them to seek cover,

Page 11: Detect Digital Image Forgeries

History of photo manipulation

Page 12: Detect Digital Image Forgeries

History of photo manipulation

• February 2004: Senator John Kerry and Jane Fonda sharing a stage at an anti-war rally emerged during the 2004 Presidential primaries as Senator Kerry was campaigning for the Democratic nomination.

Page 13: Detect Digital Image Forgeries

History of photo manipulation

Page 14: Detect Digital Image Forgeries

History of photo manipulation

• March 2004

Page 15: Detect Digital Image Forgeries

History of photo manipulation

• February 2008:

Page 16: Detect Digital Image Forgeries

History of photo manipulation

• August 2007

Page 17: Detect Digital Image Forgeries

History of photo manipulation

• November 2007

Page 18: Detect Digital Image Forgeries

Cue in Forgeries Detection

• Light Transport Difference

• Acquisition Difference

• Model Detect

Page 19: Detect Digital Image Forgeries

Detect inconsistencies in Lighting

• If the photo was composited, it’s often difficult to match the lighting conditions from individual photographs.

Page 20: Detect Digital Image Forgeries

Detect inconsistencies in Lighting

Page 21: Detect Digital Image Forgeries

Detect inconsistencies in Lighting

Page 22: Detect Digital Image Forgeries

Color Model• Assumption:

– the surface of interest is Lambertian– the surface has a constant reflectance va

lue– the surface is illuminated by a point light

source infinitely far away

Page 23: Detect Digital Image Forgeries

Image Intensity Model

• R : constant reflectance value• N(x,y) : 3 vector representing the surf

ace normal at (x ,y)• A : constant ambient light• L : surface normal

Page 24: Detect Digital Image Forgeries

Image Intensity Model

Page 25: Detect Digital Image Forgeries

Results

Page 26: Detect Digital Image Forgeries

Results

Page 27: Detect Digital Image Forgeries

Using in Forgeries Detection

Page 28: Detect Digital Image Forgeries

Detect Duplicated Image Region

• A common manipulation in tampering with an image is to copy and paste portions of the image to conceal a person or object in the scene.

Page 29: Detect Digital Image Forgeries

Forgeries Using Duplicated Image

Page 30: Detect Digital Image Forgeries

Forgeries Using Duplicated Image

• Applying PCA on small fixed size image block.– Reduce dimension representation– This representation is robust to minor va

riations in the image due to additive noise or lossy compression

• Do lexicographic sorting

Page 31: Detect Digital Image Forgeries

Results• Take 10 seconds in 512*512 image usi

ng 3 GHz processor

Page 32: Detect Digital Image Forgeries

Results

Page 33: Detect Digital Image Forgeries

Detect by Tracking Re-sample

• Processing in making forgeries often necessary to resize or rotate.

• Assume resizing by linear or cubic interpolation method.

Page 34: Detect Digital Image Forgeries

Resample

• Resample by factor of 4/3

Page 35: Detect Digital Image Forgeries

Resample

Page 36: Detect Digital Image Forgeries

Resample

Page 37: Detect Digital Image Forgeries

Resample

• Use EM algorithm to estimate

Page 38: Detect Digital Image Forgeries

Resized Estimate

Page 39: Detect Digital Image Forgeries

Rotated Estimate

Page 40: Detect Digital Image Forgeries

Rotated and Resized

• Upsampled by 15% and rotated by 5%

• Rotated by 5% and upsampled by 15%

Page 41: Detect Digital Image Forgeries

Forgery Detect

Page 42: Detect Digital Image Forgeries

PATTERN NOISE & DETECTION OF ITS PRESENCE

• Detection of digitally manipulated images based on the sensor pattern noise .

• Detection whether image take from same camera or from another region.

Page 43: Detect Digital Image Forgeries

Image Fetch Processing

Camera lense Anti-aliasing filter CFA Color interpolate

Color correction and white balence

Gamma correction and kernel filtering

Save file (JPEG compression)

Page 44: Detect Digital Image Forgeries

PATTERN NOISE & DETECTION OF ITS PRESENCE

• Most digital camera with CCD or CMOS use color filter array (CFA)

Page 45: Detect Digital Image Forgeries

PRNU• Photo-response non-uniformity noise• Dominate part of the pattern noise in n

ature images.• PNU – pixel non-uniformity : different s

ensitivity of pixel to light• Caused by stochastic inhomogenities p

resent in silicon wafer

Page 46: Detect Digital Image Forgeries

Noise Model

• xij : signal from light• ηij: random shot noise• cij: dark current• εij: read-out noise

Page 47: Detect Digital Image Forgeries

Learn PNU

• F : denoising filtering

• Training by more than 50 picture

Page 48: Detect Digital Image Forgeries

Detect• Random select n region with m

masks

• Estimate

Page 49: Detect Digital Image Forgeries

Forgery Detection Mask

Page 50: Detect Digital Image Forgeries

Forgery Detection

Page 51: Detect Digital Image Forgeries

Forgery Detection

Page 52: Detect Digital Image Forgeries

Forgery Detection

Page 53: Detect Digital Image Forgeries

Forgery Detection

Page 54: Detect Digital Image Forgeries

Reference• Luk?, J., J. Fridrich, et al. "Detecting digital image forgeries using sensor pattern nois

e." Proc. SPIE, Electronic Imaging, Security, Steganography, and Watermarking of Multimedia Contents VIII 6072: 16?9.

• Lyu, S. and H. Farid (2005). "How realistic is photorealistic?" IEEE Transactions on Signal Processing 53(2 Part 2): 845-850.

• Ng, T., S. Chang, et al. (2005). Physics-motivated features for distinguishing photographic images and computer graphics, ACM New York, NY, USA.

• Popescu, A. and H. Farid "Exposing digital forgeries by detecting duplicated image regions." Department of Computer Science, Dartmouth College.

• Popescu, A. and H. Farid (2005). "Exposing digital forgeries by detecting traces of resampling." IEEE Transactions on Signal Processing 53(2 Part 2): 758-767.

• Popescu, A. and H. Farid (2005). "Exposing digital forgeries in color filter array interpolated images." IEEE Transactions on Signal Processing 53(10 Part 2): 3948-3959.

Page 55: Detect Digital Image Forgeries

Reference• http://www.cs.dartmouth.edu/farid/r

esearch/digitaltampering/• http://www.newseum.org/berlinwall/

commissar_vanishes/vanishes.htm• http://www.cs.unc.edu/~lazebnik/res

earch/fall08/