Abstract: Digital images are highly manipulated without degrading their visual quality and resolution with advanced and easily available software’s. Copy Move Forgery (CMF) is a common manipulation procedure in digital images that involves copying the object or segment of an image and partially pasting it on another area of the same image. Many block-based detection methods are present to identify the copy move forgery in images. However, their performance got deteriorated considerably under the shape of the regions that cannot be exactly identified, and shows limited robustness based on performance metrics like precision and recall. A novel and robust algorithm is introduced in this paper to overcome that problem. The use of non-overlapping segmentation compared to the overlapping method used is to reduce the computational complexity compared to the regular blocks and irregular blocks to obtain better performance to find the forged region accurately. Here, input images are segmented using the SLIC algorithm, and to detect forgery, the key point features are extracted from each block using the Scale Invariant Feature Transform (SIFT) algorithm along with Speeded Up Robust Feature (SURF). Block matching and labelled feature point matching is used to detect the forgery from these extracted results. Block matching process is used to identify the distance between regions of the divided images. Copy move forgery region is identified by the similarities between the features of the image. Precision, Recall and F-measure of the input image are used to assess the performance of proposed scheme. It is observed that 98.97% precision is achieved with the SLIC algorithm on the MICC-F220 database.
Keywords: Image forgery, Copy-move, SLIC, SIFT, SURF, LPF
|
DOI:
10.17148/IJARCCE.2021.101208
