ROBUST COPY-MOVE FORGERY DETECTION BY FALSE ALARM CONTROL

• Author : EHRET Thibaud ehret.thibaud@gmail.com
• Licence : AGPL v3+, see agpl.txt

OVERVIEW

This source code provides an implementation (with new improvements) of "T. Ehret, ROBUST COPY-MOVE FORGERY DETECTION BY FALSE ALARM CONTROL, Arxiv 2019" available here. Please cite it if you use this code as part of your research. The structure of this code is derived from Fast Image Matching by Affine Simulation. It also uses an improved Inverse Compositional Algorithm to compute masks.

COMPILATION

The code is compilable on Unix/Linux and hopefully on Mac OS (not tested!).

Compilation: requires the cmake and make programs.

Dependencies: For image i/o we use Enric Meinhardt's iio, which requires libpng, libtiff and libjpeg. We also use StatsLib library.

Configure and compile the source code using cmake and make. It is recommended that you create a folder for building:

UNIX/LINUX/MAC:

$ mkdir build; cd build
$ cmake ..
$ make

Binaries will be created in build.

USAGE

The following commands have to be run from the build folder:

List all available options:
./main --help

There is only one mandatory input argument:

• -i the input image

Optional arguments are:

• -o the path to the output file containting the matches (string)
• -ps the patch size of the descriptor (int)
• -tau the threshold for the patch matching, also correspond to the NFA threshold when the automatic thershold is activated (float)
• -gs convert the the image to grayscale (bool)
• -auto Use the threshold defined by NFA (using tau as parameter), otherwise use the provided threshold (bool)

---

The following command compute matches that could correspond to a copy-move forgery:

./main -im ../data/forged.tif -auto true

The image in data is from the COVERAGE dataset, available here