Copy-Move Image Forgery detection via Combined Pseudo-Zernike Moment Invariant

Copy-move image forgery (CMIF) is one of the greatest threats in the field of image forensic analysis where a small region of an image is copied and pasted at other positions of the same image. This paper proposes a hybrid copy-move image forgery detection (CMIFD) approach. Initially, the SIFT algorithm is applied to the entire image and the extracted key-points are matched via 2NN process to obtain matched key-points. Furthermore, larger blocks are formed around each matched key-points followed by the computation of Combined Pseudo Zernike moment invariants (CPZMI) for each block. CPZMI feature descriptors of these blocks are simultaneously invariant to rotation, scaling, blurring and robust to noise. Consequently, comparison of CPZMI feature descriptors allows us to obtain stable matched key-points. Our proposed CMIFD technique is validated using benchmark datasets and found to exhibit better performance as compared with existing state-of-the-art techniques.