The authenticity of a digital image suffers serious threats due to the emergence of powerful digital image editing tools that easily alter the image content without leaving visible traces of such changes. In this paper, a new scheme of detection of passive splice image falsification based on the local binary pattern (LBP) and the discrete cosine transformation (DCT) is proposed. First, the chrominance component of the input image is divided into superimposed blocks. Then, for each block, LBP is calculated and transformed into frequency domain using 2D DCT. Finally, the standard deviations of the respective frequency coefficients of all the blocks are calculated and used as characteristics. For the classification, a support vector machine (SVM) is used. Experimental results in the reference splicing image falsification databases show that the detection accuracy of the proposed method is up to 97%, which is the best accuracy so far.