CoMoFoD - Image Database for Copy-Move Forgery Detection
Introduction
The aim of a digital image forensic is to determine the authenticity of digital images. One of the most common used forgery methods is a copy-move forgery, where part of an image is copied to another location in the same image. A Copy-Move Forgery Detection (CMFD) in digital images is still an open problem. Numerous algorithms have been proposed as a solution to a copy-move forgery, but there exist only few benchmarking databases for algorithms evaluation.
CoMoFoD database for a copy-move forgery detection consist of 260 forged image sets in two categories (small 512x512, and large 3000x2000). Images are grouped in 5 groups according to applied manipulation: translation, rotation, scaling, combination and distortion. Different types of postprocessing methods, such as JPEG compression, blurring, noise adding, color reduction etc., are applied to all forged and original images.
Example of image forgery in database:
 |
 |
 |
 |
| Original image | Mask | Binary mask | Forged image |