CoMoFoD — New database for copy-move forgery detection

Due to the availability of many sophisticated image processing tools, a digital image forgery is nowadays very often used. One of the common forgery method is a copy-move forgery, where part of an image is copied to another location in the same image with the aim of hiding or adding some image content. Numerous algorithms have been proposed for a copy-move forgery detection (CMFD), but there exist only few benchmarking databases for algorithms evaluation. We developed new database for a CMFD that consist of 260 forged image sets. Every image set includes forged image, two masks and original image. Images are grouped in 5 categories according to applied manipulation: translation, rotation, scaling, combination and distortion. Also, postprocessing methods, such as JPEG compression, blurring, noise adding, color reduction etc., are applied at all forged and original images. In this paper we present database organization and content, creation of forged images, postprocessing methods, and database testing. CoMoFoD database is available at http://www.vcl.fer.hr/comofod.