Image Forgery Detection using Block Matching and K-Means Clustering Techniques

The increasing prevalence of digital image forgery has led to the development of various techniques for detecting such manipulations. This study proposes a method for detecting copy-move forgery is a sort of image forgery that involves copying and pasting a segment of a picture onto another location inside the same image. Block matching and K-means clustering are the two primary phases of the suggested methodology. The picture is separated into blocks in the block matching process, and each block is compared to other blocks to identify regions that are similar. Using normalized cross-correlation, which assesses the similarity between two blocks, this stage is completed. The comparable blocks are clustered together using the K-means algorithm in the K-means clustering stage. The clusters that are produced are then examined to find areas that could have been copied and pasted. The experimental results show that the proposed technique detects copy-move forgeries with excellent accuracy even in the presence of various types of noise and post-processing. The approach is computationally efficient and may be utilized for large pictures.