Abstract: Image forgery detection is a burgeoning research field in digital forensics. As a most common way of image tampering, copy-move forgery is used to conceal objects or clone regions to produce a non-existing image scene. The target of copy-move forgery detection is to identify bigger duplicated image regions which are same or extremely similar. We reviewed several methods proposed to achieve this goal. These methods failed in detecting digital images with homogeneous texture or uniform areas and selecting the appropriate thresholds. We presented a novel adaptive thresholdbased detection algorithm for image copy-move forgery,which might be applied to the color images with homogeneous or smooth regions and identified and located forged image regions automatically if only reasonable thresholds were estimated. Experimental results on several forged images with various homogeneous or uniform regions were presented to demonstrate the effectiveness of the proposed algorithm.

Key words: Digital image forensics,Copy-move forgery,Passive detection,Adaptive