Nowadays duplication forgery is the most applicable method to make tampered images. This method copies a region of an image and pasted into another part(s) of that image. There are several methods to detect forged images. Most of them can only detect those regions which are exactly pasted into another part, but in practice the copied region is scaled or rotated before pasting to achieve the best matching with surrounding.\n In this paper we propose a method to detect duplicated regions in an image using SIFT features and then using Zernik moments. By using SIFT features we can detect duplicated regions even if they scaled or rotated, but these features cannot find flat duplicated regions. Zernik moments can solve this problem; they can find flat copied regions but failed to find scaled copied regions. So combining these two methods gives an effective approach to achieve a robust detection.

Login to Download PDF

Please login with your Paper ID and password to access the full PDF.

🔑 Login to Download