forensic detection of image manipulation using statistical intrinsic fingerprints fernando barros...
TRANSCRIPT
Forensic Detection of Image Manipulation Using Statistical Intrinsic
Fingerprints
Fernando BarrosFilipe Berti
Gabriel LopesMarcos Kobuchi
Seminar Series
MO447 - Digital Forensics
Prof. Dr. Anderson [email protected]
http://www.ic.unicamp.br/~rocha
Outline
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Outline‣ Introduction‣ System Model and Assumptions‣ Statistical Intrinsic Fingerprints of Pixel
Value Mappings‣ Detecting Contrast Enhancement‣ Detecting Additive Noise in Previously JPEG-
Compressed Images‣ Conclusion
Introduction
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Nowadays...
‣ In recent years, digital images have become increasingly prevalent through society.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
© Daily Stormers (www.dailystomers.com)
© R
aíz d
a V
ida (w
ww
.raizd
avid
a.co
m.b
r)
© http://topwalls.net/grand-canyon-united-states-2/
Real Pictures
© www.wallpea.com
*2013 Seminar Series – Digital Forensics (MO447/MC919)
© Fan Pop (www.fanpop.com)Digital Images
© johnnyslowhand.deviantart.com
© www.highqualitywallpapers.eu
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Fake Pictures!
© www.epicfail.com
© www.hoax-slayer.com
©ig
ossip
.com
© fashmark.files.wordpress.com
*2013 Seminar Series – Digital Forensics (MO447/MC919)
© www.vidrado.com
Fake Pictures?
© Veja (http://veja.abril.com.br/)
© Telegraph (www.telegraph.co.uk)
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Fake Pictures?© http://www.buzzfeed.com/tomphillips/22-viral-pictures-that-were-
actually-fake
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Fake Pictures?
© http://10steps.sg/inspirations/photography/70-strange-photos-that-are-not-photoshopped/
*2013 Seminar Series – Digital Forensics (MO447/MC919)
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Consequence
‣ At present, an image forger can easily alter a digital image in a visually realistic manner.
‣ As a result, the field of digital image forensics has been born.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
State Of The Art
‣ Identification of images and image regions which have undergone some form of manipulation or alteration
‣ No universal method of detecting image forgeries exists
‣ Different techniques, with their own limitations
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Some techniques
‣ Lighting Angle Inconsistencies
‣ Inconsistencies in chromatic aberration
‣ Absence of Color Filter Array (CFA) interpolation induced correlations
‣ Classifier based approaches
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Fingerprints
‣ Most image altering operation leave behind distinct, traceable “fingerprints” in the form of image alteration artifacts
‣ Because these fingerprints are often unique to each operation, an individual test to catch each type of image manipulation must be designed
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Some works with fingerprints
‣ Resampling
‣ Double JPEG compression
‣ Gamma correction
*2013 Seminar Series – Digital Forensics (MO447/MC919)
This work‣ Pixel value mapping leaves behind statistical
artifacts which are visible in an image’s pixel value histogram
‣ By observing the common properties of the histogram of unaltered images, it’s possible to build a model of an unaltered image’s histogram
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ A number of operations are in essence pixel value mapping, it’s proposed a set of image forgery detection techniques which operate by detecting the intrinsic fingerprint of each operation
This work
System Model and
Assumptions
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Digital image‣ In this work, digital images created by using an
electronic imaging device to capture a real world scene
‣ Each pixel is assigned a value by measuring the light intensity reflected from a real world scene
‣ Inherent in this process is the addition of some zero mean sensor noise which arises due to several phenomena (shot noise, dark current, etc)
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ For color images, it is often the case that the light passes through a CFA so that only one color component is measured at each pixel location in this fashion
‣ In that case, the color component not observed at each pixel are determined through interpolation
‣ At the end of this process the pixel values are quantized, then stored as the unaltered image
Color image
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ h(l) can be generated by creating L equally spaced bins which span the range of possible pixel values
‣ Tabulate the number of pixels whose value falls within the range of each bin
‣ Gray levels values in P = {0, … , 255}, Color values in P³
‣ Pixel value histogram uses 256 bins
Histogram
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histograms‣ None of the histograms contains sudden zeros
or impulsive peaks
‣ Do not differ greatly from the histogram’s envelope
‣ To unify these properties, pixel value are described as interpolatably connected
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Interpolatably connected‣ Any histogram value h(l) can be aproximated
by ĥ(l)
‣ Each value of ĥ has been calculated by removing a particular value from h then interpolating this value using a cubic spline
‣ Little difference from h and ĥ
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Fig. 1. Left: Histogram of a typical image. Right: Approximation of the histogram at left by sequentially removing then interpolating the value of each histogram entry.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
System Model‣ To justify this model, a database of 341
unaltered images captured using a variety of digital cameras
‣ Obtained each image’s pixel value histogram h and its approximated histogram ĥ
‣ The mean squared error between both along with the signal power of h to obtain an SNR~30.67dB
Statistical Intrinsic
Fingerprints of Pixel Value Mappings
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Pixel Value Mapping
‣ A number of image processing operations can be specified entirely by a pixel value mapping
‣ Leave behind distinct, forensically significant artifacts, which we will refer as intrinsic fingerprint
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Intrinsic Fingerprint
‣ Intrinsic Fingerprint
‣ Original: x; Tampered: y
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Discrete Fourier Transform
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Example: Histogram
Synthesized Image
Real World Image
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Example: DFT
Synthesized Image
Real World Image
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Example: Frequency Domain
Synthesized Image
Real World Image
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Example: Frequency Domain
Synthesized Image
Real World Image
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ When examining a potentially altered image, if the histogram of unaltered pixel values is known, the tampering fingerprint can be obtained using
Intrinsic Fingerprints
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ In most real scenarios, one has no a priori knowledge of an image’s pixel value histogram, thus the tampering fingerprint cannot be calculated
But...
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ It’s possible to ascertain the presence of a tampering fingerprint by determining identifying features of a mapping’s intrinsic fingerprint
‣ Searching for their presence in the histogram of the image
However
Detecting Contrast
Enhancement
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Overview
‣ Contrast Enhancement operations seek to increase the dynamic range of pixel values within images.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Detection of Globally Applied Contrast Enhancement‣ Usually nonlinear mappings
‣ Consider only monotonic pixel value mappings.
‣ Thereby, disconsidering simple reordering mappings.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Two significant mappings
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Euclidian norm increases!
‣ The energy of the DFT as well
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Therefore, all contrast enhancement mappings result in an increase in energy.
‣ This energy is related to the intrinsic fingerprint.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast‣ Expected DFT’s to be strongly low-pass signal.
‣ Therefore, the presence of energy in the high frequency regions is indicative of contrast enhancement.
‣ Contrast enhancement will cause isolated peaks and gaps in the histogram.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast‣ Saturation Case
*2013 Seminar Series – Digital Forensics (MO447/MC919)
http
://w
ww
.fun
nyju
nk.
com
/funn
y_p
ictu
res/
3306
291/
Ho
usto
n+w
e+
have
+a+
pro
ble
m/
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Solution
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast‣ Measuring the energy
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Defining the best c with 244 images e Np = 4
‣γ = 1.1
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Results
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Database of 341 unaltered images, taken in different resolutions and light conditions
‣ The green color layer created the grayscale images.
‣γ ranging from 0.5 to 2.0
‣ 4092 grayscale images
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Globally Applied Contrast
‣ Np = 4 e c = 112
‣ Pd of 99% at a Pfa approximately of 3% or less
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ Defined as applying a contrast mapping to a set of contiguous pixels within an image.
‣ Can identify cut-and-paste forgeries.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ To detect it, the image is divided in smaller blocks and the global technique is applied to the blocks.
‣ Who small(and big!?) are these blocks?
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ Test the 341 unaltered images and use γ ranging from 0.5 to 0.9.
‣ Blocks of size 200x200, 100x100, 50x50, 25x25, and 20x20
*2013 Seminar Series – Digital Forensics (MO447/MC919)
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ The contrast enhancement can be reliably detected using testing blocks sized 100x100 pixels with a Pd of at least 80% in every case at a Pfa of 5%.
‣ When γ ranged from 1.0 to 2.0, the Pd was of 95% at a Pfa of 5%
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ In order to test the copy-and-paste, Photoshop was used to create the image (c).
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ The image was divided in 100x100 pixel blocks and tested local contrast enhancement on the red(d), green(e) and blue(f) color layers.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ The image was divided in 50x50 pixel blocks and tested local contrast enhancement on the 3 the color layers.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Locally Applied Contrast
‣ Applying a detection criteria.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ Histogram equalization effectively increases the dynamic range of an image’s pixel values by subjecting them to a mapping such that the distribution of output pixel values is approximately uniform.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ In order to identify it, we calculate the “uniformity” of the histogram.
‣ The process will introduce zeros into an image’s pixel value histogram, so mean absolute differences and mean square differences won’t work.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
[2]
http
://w
ww
.fun
nyju
nk.
com
/funn
y_p
ictu
res/
3306
291/
Ho
usto
n+w
e+
have
+a+
pro
ble
m/
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ For histogram equalized saturated images, the location of the impulsive component is often shifted.
‣ Suppose that the number of pixels in the lowest bin is greater than 2N/255.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ If the lowest l such that h(l) > 0 is greater or equal to 1 and h(l) > 2N/255, the image is identified as saturated.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ Test the 341 unaltered images and the 341 histogram equalized images.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ Analyze the frequency domain.
‣α(k) is a weighting function used to deemphasize the
high frequency regions in H(k)
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ Best conditions using α1 (k) was with r1 = 0.5, obtaining Pd of 99% with a Pfa of 0.5% and a Pd of 100% with a Pfa of 3%.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Histogram Equalization
‣ Test 2046 images.
‣ r2 = 4
‣ Pd of 100% with a Pfa of 1%.
Detecting Additive Noise in Previously
JPEG-Compressed
Images
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise‣ Additive noise can be used to mask previous
modifications to images.
‣ Previous techniques has dealt with detection of localized fluctuations of SNR in an image.
‣ Fail on detection of globally added noises.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise‣ This technique applies a predefined mapping with a known fingerprint to a potentially altered image.
‣ If some noise was intentionally added, then an identifying feature of this fingerprint will be absent.
‣ We’ll be able to detect the presence of an additive noise if application of mapping does not introduce a fingerprint with this feature.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Scale and Round mapping‣ For additive noise detection it’ll be used scale
and round mapping:
‣ And the set
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Scale and Round mapping‣ Cardinality of UC(v) is periodic in v with period
p.
‣ So, the intrinsic fingerprint of scale and round operation will contain a periodic component with period p.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ JPEG compression/decompression schematics
© Compressed Examples by JISC Digital Media. All files © University of Bristol, 2009
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ So, if a monotonically increasing mapping is
applied to any color layer in the YCbCr color space, that mapping’s fingerprint will be introduced into the histogram of the color layer value.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ Final stage of JPEG decompression: pixel
transformation from YCbCr to RGB, mathematically described by this equation:
‣ Values less than 0 is set to 0 and greater than 255 is set to 255.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ Defining:
‣ Detection of additive noise can be formulated as an statistical hypothesis testing problem:
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ The fingerprint left by the mapping:
‣ helps to rewrite both hypothesis as:
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ Under hypothesis H0 , zi can be expressed as:
‣ The term round(cxi) dominates the behavior of zi and, so, the number of distinct xi values mapped to each zi value will occur in a fixed periodic pattern.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ This will result in a periodic pattern discernible
in the histogram, which corresponds to the intrinsic scale and round mapping.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ Under hypothesis H1 , zi has a different
behavior:
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario‣ This hypothesis leads to 3 additional terms
containing scale and round mapping, each with their own scaling constant.
‣ If this constants and the original scaling has no common period, no periodic pattern will be introduced into the histogram, as can be observed in the figures.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Hypothesis Testing Scenario
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise Detection Images‣ Detection of the addition of noise to a
previously JPEG-compressed images is the same as detection of the periodic fingerprint within the normalized histogram.
‣ This detection is well suited for frequency domain and produces peaks with arbitrary location.
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise Detection Images‣ Applying DFT in the normalized and pinched-off
histogram we obtain Gzi, it is possible to measure the strength of the peak introduced into it:
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise Detection Images
‣ Then it is used the following decision rule to determine presence of additive noise:
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ 227 unaltered images from 4 different digital cameras from unique manufacturers.
‣ Diversity of JPEG-compressed images using camera’s settings.
‣ Set of altered images created by decompression and addition of unit variance Gaussian noise to each pixel value.
Additive Noise Detection Images - 1st performance test
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ All altered saved with original images resulting in a DB of 554 images.
‣ Pd = 80% @ Pfa = 0,4% (if Pfa <= 6,5% Pd goes to nearly 99%).
Additive Noise Detection Images - 1st performance test
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise Detection Images - 1st performance test
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ 244 JPEG-compressed images at different quality. ratios.
‣ Q=90, 70, 50 and 30.
‣ Again, unit variance Gaussian noise added.
Additive Noise Detection Images – 2nd performance test
*2013 Seminar Series – Digital Forensics (MO447/MC919)
‣ For images with Q >= 50.
‣ Pd = 99% @ Pfa = 3,7%.
Additive Noise Detection Images – 2nd performance test
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Additive Noise Detection Images – 2nd performance test
Conclusions
*2013 Seminar Series – Digital Forensics (MO447/MC919)
Conclusions‣ Statistical Intrinsic Fingerprints of Pixel
Value Mappings.
‣ Detecting Contrast Enhancement
‣ Detecting Additive Noise in Previously JPEG-Compressed Images
References
*2013 Seminar Series – Digital Forensics (MO447/MC919)
References1. A. Swaminathan, M.Wu, and K. J. R. Liu, “Digital image forensics via intrinsic fingerprints,” IEEE
Trans. Inf. Forensics Security, vol. 3, no. 1, pp. 101–117, Mar. 2008.
2. http://www.jiscdigitalmedia.ac.uk/guide/file-formats-and-compression/
Thank You!Obrigado!