Copy-move forgery detection system through fused color and texture features using firefly algorithm

Abstract

�Abstract: Copy-Move Forgery Detection (CMFD) is an established process to detect copy-move tampered regions in digital images. Several CMFD algorithms based on image transform, color and texture features are available in the literature. Detection of the tampered regions depends on the superiority of the feature vector. Hence, an efficient passive approach is proposed in which color and texture features are fused to form an improved feature vector. Firefly Algorithm (FA) is explored to obtain the nonlinear relationship between color and texture features. These Optimal Weighted Color and Texture Features (OWCTF) are used for detection of forged images and later localization is performed to detect the tampered regions in the forged image. The detection performance of the proposed method is evaluated on CASIA and CoMoFoD databases and the classification accuracy of 95.5% and 97% is achieved respectively. Similarly, performance evaluation of localization phase is also carried out. Simulation results demonstrate that the proposed method overtakes some of the existing methods in terms of detection and localization results. It is witnessed that proposed method is capable to detect and localize the tampered regions in the presence of signal processing attacks.