Doctored JPEG image detection based on double compression features analysis

Identifying the authenticity and integrity of digital images becomes increasingly important in digital forensics. In this paper, we focus on JPEG images and propose an effective method for detecting doctored images. We first investigate the statistical characteristics of DCT coefficients based on a recompression files sets, and analyze the differences of double compression effect between doctored and non-doctored region in a doctored image. We then extract the DCT coefficients histograms of each block in doctored images and represent them as feature vectors. We identify the location of doctored region by using SVM classification for evaluating the feature vectors. Experimental results demonstrate that the proposed method can efficiently detect and automatically locate doctored regions on different forgeries with low computational complexity.