Techn. Fakultät Website deprecated and outdated. Click here for the new site. FAU-Logo

Copy-Move Forgery Detection

Copy-Move Forgery Detectors and Ground Truth Generator

This software package contains the core components (code, some scripts) for our paper "An Evaluation of Popular Copy-Move Forgery Detection Approaches" by V. Christlein, C. Riess, J. Jordan, C. Riess and E. Angelopoulou.

The software consists of a single binary. Subcommands can be passed as the first parameter to either create ground truth for Copy-Move Forgery Detection (CMFD), or to apply a CMFD detector to an image.

To build the code, you require a) a recent version of OpenCV, b) a recent version of the boost library and c) cmake. The ground truth generator comes with a couple of perl scripts, which are not critical to the functionality of the software. We tested the code on a debian squeeze Linux; however, no platform-specific code is used in the C++ part. With some fiddeling in (or a replacement of) the build scripts, it should build smoothly on Windows as well.

The package contains a couple of README_*.txt files. Start in the base file (README.txt) for pointers to the remaining files. For instance, build instructions are contained in README_build.txt. In case you have questions, feel free to write to Opens internal link in current windowChristian Riess or Vincent Christlein.

Note that the code is provided as-is. We aimed to carefully implement prior work. However, it may well be that we missed a good trick to improve the performance. If you find a bug, feel free to drop us an email, or directly send us a patch for the respective method.

The code is free to use for research purposes. If you publish work containing results from the code, please cite our paper.

Initiates file downloadDownload the code (v. 1.3, including bugfixes by Li Jian, Seung-Jin Ryu and John Edgar Vargas; thanks also to Renata Ghisloti Duarte de Souza for a fix in the build description).

Opens internal link in current windowTo the project page

Opens internal link in current windowTo the dataset (<- also required for ground truth generation)