Post on 21-Apr-2018
transcript
Ranking of manipulated images in a large set using Error Level Analysis
Daan Wagenaar & Jeffrey Bosma
University of Amsterdam
In cooperation with the Netherlands Forensic Institute
12-02-12
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
2
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
3
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Image Manipulation
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
4
¤ Examples ¤ Red Eye removal
¤ Brightness enhancements
¤ Sharpening
¤ …
¤ Most interesting manipulations ¤ Internal copy & move
¤ External copy & move
5
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Stalin with Yezhov (original)
Stalin without Yezhov (manipulated)
Object removal
6
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Object appearance modification
Katie Couric (original)
Slimmed Body (manipulated)
7
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Object addition
Holding an iPhone (original)
Holding a BlackBerry (manipulated)
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
8
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Research Question
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
9
¤ Problem: ¤ A set of images as part of evidence
¤ An expert manually inspects each image for manipulations
¤ Time consuming process in a large set of images
v Can the Error Level Analysis technique be used to rank a set of images according to potentially present image manipulation?
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
10
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Error Level Analysis (ELA)
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
11
¤ A technique for detecting image manipulations
¤ Uses properties of lossy image format
¤ Compares error caused by compression to a certain quality level
¤ An example: ¤ Initial image at a quality level of 95%
¤ ELA resaves this image at a certain quality level (e.g. 95%)
¤ Compression introduces error
¤ Compare error of initial and resaved image
¤ Manipulated areas will have a different level of error
¤ Differences are visibly expressed by brightness in a third image
12
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Original image ELA @75%
ELA @ 85% ELA @ 95%
13
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Manipulated image ELA @75%
ELA @ 85% ELA @ 95%
14
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
¤ Limitations ¤ False positives can be caused by:
¤ Sharp contrast, well-defined patterns
¤ Recoloring, such as brightening, pallet skew, ...
¤ False negatives can be caused by:
¤ Low resolutions
¤ Scaling
¤ Low quality
¤ Image scanning from other sources
¤ Extremely skilled artists
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
15
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Methodology
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
16
¤ Method 1: Average RGB values per block
¤ Method 2: Block to block comparison
¤ Method 3: Colored pixels ratio
¤ Method 4: Highest luminance value of the brightest pixel
¤ Method 5: Average luminance value of the 64 brightest pixels
¤ Method 6: Average luminance value of the brightest block
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
17
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Experiments
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
18
¤ Goal
¤ Proof of concept
¤ Dataset of 300 images ¤ 100 images with Canon PowerShot A630
¤ 100 images with iPhone 4
¤ 100 images with Samsung Digimax S500
¤ 30 manipulated images
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
19
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Results
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
20
0 50 100 150 200 250 300Rank
0
20
40
60
80
100
Man
ipul
ated
imag
es fo
und
(%)
Method 3 (Colored pixels ratio)Method 4 (Highest luminance value of the brightest pixel)Method 5 (Average luminance value of the 64 brightest pixels)Method 6 (Average luminance value of the brightest block)
Rankings with ELA at 75%
21
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
0 50 100 150 200 250 300Rank
0
20
40
60
80
100
Man
ipul
ated
imag
es fo
und
(%)
Method 3 (Colored pixels ratio)Method 4 (Highest luminance value of the brightest pixel)Method 5 (Average luminance value of the 64 brightest pixels)Method 6 (Average luminance value of the brightest block)
Rankings with ELA at 85%
22
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
0 50 100 150 200 250 300Rank
0
20
40
60
80
100
Man
ipul
ated
imag
es fo
und
(%)
Method 3 (Colored pixels ratio)Method 4 (Highest luminance value of the brightest pixel)Method 5 (Average luminance value of the 64 brightest pixels)Method 6 (Average luminance value of the brightest block)
Rankings with ELA at 95%
23
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Manipulated image ELA @75%
ELA @ 85% ELA @ 95%
24
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Manipulated image ELA @75%
ELA @ 85% ELA @ 95%
25
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Manipulated image ELA @75%
ELA @ 85% ELA @ 95%
26
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
Original image ELA @75%
ELA @ 85% ELA @ 95%
27
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
ELA @75%
ELA @ 85% ELA @ 95%
Manipulated image
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
28
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Conclusion
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
29
¤ Most effective method
¤ Limitations of ELA directly affect developed methods
¤ Detectable manipulation techniques
v Can the Error Level Analysis technique be used to rank a set of images according to potentially present image manipulation? ¤ Yes, it is possible albeit not very reliable.
Agenda
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
30
¤ Image Manipulation
¤ Research Question
¤ Error Level Analysis
¤ Methodology
¤ Experiments
¤ Results
¤ Conclusion
¤ Further Research
¤ Questions
Further Research
12-02-12 Ranking of manipulated images in a large set using Error Level Analysis
31
¤ Alternative to ELA
¤ Combine different rankings
¤ Different methods