Copy-Move Forgery Detection Algorithm based on Feature Point Clustering

Aiming at the high time complexity of the feature matching stage of the current copy-move forgery detection algorithm, an image copy-move forgery detection algorithm using structure tensor and HSV color model to cluster feature points is proposed. First, cluster the SIFT feature points based on the structure tensor, and divide all feature points into flat feature points, edge feature points, and corner feature points, which are divided into 3 clusters; Then, based on the clustering method of HSV color model, the feature points are divided into 63 clusters. Finally, feature matching is carried out in each cluster, which makes full use of the similarity of texture and color between the source region and the tampered region, effectively reduces the time of feature matching and improves the efficiency of the algorithm. Experimental results show that the proposed algorithm can effectively detect tampered areas, has a greater advantage in matching time, and has good robustness.