Abstract: Through analysis on the characteristic of image splicing，a new image splicing blind detection approach was proposed. It extracts the gray level cooccurrence matrix（GLCM）based on the thresholded DCT coefficients of the image. The gray level cooccurrence matrixes along the four directions serve as feature vector for image splicing detection. Support vector machine（SVM）is utilized as the classifier in the proposed method. The experimental results indicate that this new scheme can achieve the accuracy of 91.2% and 98.5% on gray image dataset of Columbia University and color image dataset of Chinese academy of sciences (CAS) respectively.

Key words: image splicing detection； multi block discrete cosine transform (DCT)； gray level cooccurrence matrix； support vector machine (SVM)