Detecting Resized Double JPEG Compressed Images – Using Support Vector Machine

摘要: Since JPEG is the most widely used compression standard, detection of forgeries in images necessary. In order to create a forged image, image usually loaded into photo editing software, manipulated and then re-saved as JPEG. This yields double artifacts, which can possibly reveal forgery. Many techniques for compressed have been proposed. However, when resized before second step, blocking artifacts first are destroyed. Therefore, reported detecting do not work this case. paper, we propose technique (called RD-JPEG) images. We identify features that discriminate RD-JPEG from use Support Vector Machines (SVM) classification tool. Experiments with many different quality scaling factors indicate our works well.