Abstract: To deal with image tampering operation using seam carving, a forensic method based on LBP and Markov features is proposed. Firstly, the LBP operator is exploited to convert the image from pixel domain to LBP domain. After that, the 2D JPEG matrix is extracted by JPEG compression. Then its first-order difference matrices are calculated in the horizontal, vertical, diagonal and minor diagonal directions respectively. Finally, the Markov features are gained from the difference matrices in each direction and sent to the SVM to identify whether an image is original or suffered from seam carving. The experiment result shows that the proposed algorithm is superior to the traditional Markov feature algorithm and other existing seam carving detection algorithms. Especially when tampering is small scaling seam carving, the detection efficient will obviously improved.