Detection of Shifted Double JPEG Compression using Markovian Transition Probability Matrix

Copy-paste forgery is a very common type of forgery in JPEG images. The tampered patch has always suffered from JPEG compression twice with inconsistent block segmentation. This phenomenon in JPEG image forgeries is called the shifted double JPEG (SD-JPEG) compression. Detection of SD-JPEG compressed blocks can make crucial contribution to detect and locate the tampered region. However, the existing SD-JPEG compression detection methods cannot achieve satisfactory results especially when the image block size is small. In this paper, an effective SD-JPEG compression detection method based on Markovian transition probability matrix is proposed. Statistical artifacts are left by the SD-JPEG compression among the elements of the JPEG 2-D arrays. Difference JPEG 2-D arrays generated along four directions (i.e. horizontal, vertical, main diagonal and minor diagonal) are utilized to enhance them and then thresholded by a predefined threshold for reducing computational cost. Markovian transition probability matrix is used to model the difference JPEG 2-D arrays in order to utilize the second order statistics. All the elements of these transition probability matrices are served as features for SD-JPEG compression detection. Support Vector Machine (SVM) is employed as the classifier. Experimental results demonstrate the efficiency of the proposed method.