Copy Move Forgery Detection through Graph Neighborhood Degree

摘要

In this digital world, data and image are two main components which can be obtained with ease on the internet. This aspires to copy-move parts of an image from one region to another on the same image to give a new fallacious perspective. In this paper we have presented a novel copy move forgery detection scheme using graph neighborhood degree. At first, non-overlapping sub-blocks are taken of an image to form a codebook against a threshold. The corresponding pixels of the obtained codebook are processed to embed data at LSB of the pixels in the block. When this image is forged, we again compute the bits of each block to be embedded as previous and match to the LSB of that block. If unmatch ensues, the image is forged. Subsequently, the experiments are conducted on three standard databases to detect the efficiency of the proposed algorithm on copy-move forged images. Also, some tampering detection schemes are carried out to show the robustness and coherence of the proposed algorithm.