Image Copy-Move Forgery Detection Based on NSST

This paper proposes an image block feature extraction method based on Nonsubsampled Shearlet Transform (NSST) transform combined with Fast Walsh Hadamard Transform (FWHT) instead of DCT. It can improve the algorithm's detection speed without reducing the detection accuracy. In addition, a Multi-Hop Jump(MHJ) is used to improve the matching efficiency. It will skip some unnecessary matching blocks. Experiments show that compared with the methods of Fridrich, Popescu, Kulkarni, the accuracy of this method is partially improved, and the computational efficiency is greatly improved.