We are happy to inform that our article “Behavior Knowledge Space-Based Fusion for Copy–Move Forgery Detection” has been published in IEEE Transactions on Image Processing, vol. 25, no. 10, pp. 4729-4742, 2016. The authors tackle the problem of copy-move image tampering, which can be used for misleading the opinion forming process of the general public.
The article goes beyond traditional forgery detectors which either takes image patches containing raw pixels and, by lexicographical sorting and thresholding, similar patches are found in the image or uses similarity of points of interest to find copied and pasted regions. A more interesting approach would combine these approaches through fusion techniques. Thus, the work deals with the limitations of fusing approaches by designing a robust and efficient Behaviour Knowledge Space representation more appropriate for copy-move detection, modeling the problem as a conditional probability estimation problem instead.
Experimental results are shown on complex datasets, comparing the proposed techniques with a set of copy-move detection approaches and other fusion methodologies in the literature.
A pre-print version is available here.