Logo

Establishment of a benchmarking image dataset of copy-move text forgeries

DegreeBachelor
StatusClosed
Supervisor(s)Svetlana Abramova, MSc

Description

The detection of copy-move forgeries has been studied extensively in the image forensics literature, however all known methods were designed and evaluated only with digital images depicting natural scenes. This thesis addresses the problem of copy-move tampering with scanned versions of printed documents.

In particular, the goal of this thesis is to develop a toolbox for the automatic creation of copy-move forgeries of scanned text images by means of OCR techniques, and to generate an extensive dataset of forged documents of different scanner resolutions and scenarios. The toolbox has to enable the following processing operations: JPEG pre- or post-compression, rotation of glyphs, noise addition, or combination of these. In addition to generating a copy-move text forgery, it has to deliver a corresponding ground-truth binary map of copy-move and original pixels.

References

  • Christlein, V., Riess, C., Jordan, J., Riess, C., and Angelopoulou, E. An Evaluation of Popular Copy–Move Forgery Detection Approaches. IEEE Transactions on Information Forensics and Security, 7, 6 (2012), 1841–1854.
  • Abramova, S. and Böhme, R. Detecting Copy–Move Forgeries in Scanned Text Documents. In Proceedings of IS&T Electronic Imaging: Media Watermarking, Security, and Forensics 2016. San Francisco, CA, 2016. [PDF]