ITE513 FORENSIC INVESTIGATION

Topic 10
Image Forgery
Topic Reading

• Chapter 20 of text.

• Additional readings on Interact2.

Image Forgery Detection
Images

• Are nearly as old as computing itself.

• Used everywhere for many varied reasons.

• News, entertainment, information, memories, evidence…etc.

• Very simple to edit and manipulate in obvious, and covert, ways.

• Many image manipulation packages.
• Popular for aesthetics often!

• Seeing is no longer believing.

Image Forgery

• So simple to do using a variety of techniques.

• Detecting image forgery uses analysis of statistical characteristics and

formation mechanisms.

• Image processing has two main focuses:

• Improving image quality: for us.
• Processing for storage, transmission or display: for digital systems.
Image Concepts
Image Types
Image Types
The Math

• There is a whole heap of math in the text (Section 20.1.3, p.512-517).

• Read if you so desire but not necessary for our purposes!
• Even if it was…I couldn’t help you out.
Detecting Forgery

• An essential role in forensics because images are so important in building

a story or “truth”.
– Social.
– Political.
– Legal.
– Educational.
– Marketing.
– Historical.
– Propaganda.
– “fake news”.
– Other……
Detecting Forgery
Detecting Forgery

Source: https://www.sciencedirect.com/science/article/pii/S266591072030061X
Detecting Forgery

• Reveals the fundamental weakness of images.

• Verification of authenticity is vital in many contexts.

• Huge technology advances in this area in recent times.

• Investigators need to keep pace.

• https://www.swgit.org/ - was one such organisation.
• Now defunct.
• https://www.swgde.org/ - still current ☺
– https://www.swgde.org/documents/published
Tampering Techniques

• Copy-Move.

Source: https://link.springer.com/chapter/10.1007/978-3-030-38752-5_2
Tampering Techniques

• Image-Splicing.
Tampering Techniques

• Removal.

Source: https://cmte.ieee.org/futuredirections/2018/08/23/6953/
Protecting against Forgery

• Two main technology “types”:

• Active defence: Watermarking, Digital signature.

• Passive blind detection.

Watermarking
Digital Signature
Passive Blind

• Most people can’t be bothered with watermarks or digital signatures.

• Complex, time consuming.

• Passive-blind is simple and quick.

• Identifies copied sections from pixels in the original.
• Examines inconsistencies in statistics of natural images.
• Questions which device/s was/were used to capture the image.
Copy-Move Forgery Detection

• Simple and effective form of detection.

• Block-based.
• Compatible with many feature extraction techniques.
• Looks for similarity in multiple areas of an image.

• Keypoint-based.
• Examines distinctive features in an image.
– Corners, blobs, edges etc.
– Classifies and matches.
Device-based Forgery Detection

• Digital images can be captured via various devices.

• Cameras, scanners etc.

• If the capturing device can be identified this can be used to verify the
authenticity of the image.

• Uses:
• Sensor noise.
• Chromatic aberration.
• Colour filter array.