Detecting Asymmetric Cropping Based on Camera Calibration

A new forensics method is proposed to detect asymmetric cropping. According to the method, the inconsistency of the principal point is used as evidence of cropping, and the camera calibration method is introduced to estimate the principal point. The camera calibration model is improved to meet the requirements of image forensics, in the new model, the target objects are selected from those regular figures on different planes, so the principal point coordinate can be estimated from a single image without modeling and measuring the target objects. Experiments show that the proposed method can identify asymmetric cropping of the image effectively.