Abstract: Image splicing is exceptionally normal and central in picture altering. Along these lines, picture splicing detection has pulled in increasingly more consideration as of late in digital image forensics. grey pictures are utilized straightforwardly, or colour pictures are changed over to grey pictures before be handled in past picture splicing detection algorithms. Be that as it may, most forged pictures are can be color or grey pictures. So as to utilize the grey data in pictures, a classification algorithm is upgraded which can utilize grey pictures with Spatial, DCT and DWT features directly. In this paper, an algorithm dependent on Markov chain with spatial, DCT and DWT area is proposed for picture splicing detection. As a matter of first importance, grey data is generated from blocked pictures to build DCT and DWT in an entire way, and the DCT and DWT coefficients of blocked pictures can be obtained. Furthermore, the expended Markov features created from the Transition probability matrix in spatial, DCT and DWT area can catch the intra-block, yet additionally the Inter block correlation between's blocked DCT and DWT coefficients. Then we use PCA for reducing the dimensionality of pictures and enhancing the correlation among pixels. At long last, ENSEMBLE classifier is used to classify the Markov feature vector. The final results show that the proposed algorithm not just utilize grey data of pictures, yet in addition can yield significantly better detection results in contrasted to previous work for splicing detection methods applied on same equivalent dataset
Keywords: Image DCT, DWT, Markov Chain, PCA, Ensemble Classifier.
| DOI: 10.17148/IJIREEICE.2019.7419