IMAGE COPY-MOVE FORGERY DETECTION BASED ON SPATIAL DOMAIN

Abstract

A prominent trend in information deception is the manipulation of digital images. One kind of manipulation is 
copy-move forgery. Whether active (with prior information of the source picture) or passive (without), image forgery is the 
study of identifying image altering. Even after the copied area has been rotated in the spatial domain, our proposed approach 
for recognizing Copy-Move regions is efficient and quick. The suggested approach uses parallel block comparisons to speed 
up the blocking matching mechanism. The first step in feature extraction is to split the picture into overlapping blocks of a 
given size. The blocks are grouped into distinct clusters using the k-means clustering algorithm. A similarity measure is 
computed between each adjacent block to ascertain their degree of similarity after the feature vectors of every cluster block 
are lexicographically sorted using radix sort. Even after an image has been altered by jpeg compression, rotation, or 
smoothing conditions, the suggested approach may effectively recognize the duplicated parts, according to the testing 
findings. When compared to other efforts, the suggested solution cut processing time by as much as 75%.  
Blind image forensics, picture falsification detection, copy-move forgery, image manipulation, and image tampering are all 
related terms.