Efficient image splicing detection algorithm based on markov features

Image splicing is one of the most common methods for digital image tampering. In this paper, an efficient Markov features based algorithm is proposed for image splicing detection. The proposed algorithm first extracts two types of Markov features, coefficient-wise Markov features and block-wise Markov features in the discrete cosine transform (DCT) domain. The former are obtained by exploiting correlations between consecutive coefficients and the latter are computed by exploiting coefficient correlations between adjacent blocks. Then, a feature vector is obtained by combining these two Markov features and it is fed into support vector machine (SVM) for the classification of authentic and spliced images. The experimental results show that the proposed method not only achieves much higher detection accuracy but also reduces the total running time significantly in comparison with state-of-the-art methods.