Digital Image Forensic - Methods: Overview

Passive blind image forensics

Without any access to an original

Format based approaches

  • Source identification
  • Devicetype identification
  • Device Identifikation
  • RAW-Tool identification
  • JPEG analysis
    • Reconstruction of a compression history
    • Double compression detection (aligned or non-aligned)
    • Quality estimation
    • Thumbnail comparison
  • Analysis of metadata

Hardware based approaches

  • Dust
    • Source camera identification
  • Lens
    • CA-based splicing detection
    • Distortion based focal length estimation
  • Sensor
    • CFA-type recognitions or estimation
    • CFA-pattern based splicing recognition
    • PRNU based device identification


Physics based approaches

  • 3D-reconstruction
  • Astronomical inconsistency detection (Sun and moon position or size)
  • Geolocation estimation
  • Perspective inconsistency detection
  • Light source estimation
  • Photogrammetry
    • Projective geometry and orthorectification
    • Reverse projection
  • Single view metrology

Soft- and firmware based approaches / Pixel-based approaches

  • Splicing detection
  • Cloning or copy-move forgery detections
  • Resampling detection


Statistics based approaches

  • Noise inconsistency detection
  • Histogram aberration detection
  • Blind source separation



Passive semi-blind image forensics

Without any access to an original, but to other images taken at the same place or the same time

Comparative methods

  • Comparison with images of a series


  • Comparison with third party images

Passive non-blind image forensics

With access to an original or raw data

Comparative methods

  • Comparison with single image versions or raw-data


  • Comparison of lage image data sets

Active image forensics


  • Tasks of Hash-Encryption or Watermarking will be handled by cooperation partners



Signal and image processing


Signal reconstruction

  • Blind deconvolution
  • Superresolution
  • Signal reconstruction
  • Image calibration
  • Information extraction



Machine learning


Face reconition

  • Convolutional Neural Network (CNN) / Support Vector Machine (SVM) classifier based