Photographic composite detection using scene perspective constraints

Abstract

Photographic composite is a common image forgery which inserts visually plausible composite into target image. However, the size relationship of multiple objects in real world might be distorted since it is difficult to insert a foreign object into the target image with proper size. To detect photographic composite forgery, a blind detection approach is proposed by exploiting scene perspective constraint. It determines the height ratio of two objects in an image without any knowledge of camera parameters. That is, the height ratio is found merely by the planar homography of reference plane on which both objects of interest are situated and the vanishing point of the direction perpendicular to the reference plane. If the relationship of relative size is not appropriate, candidate image is identified as a forged image. Experiment results prove that the proposed method is effective, even when candidate images have been down-sampled or compressed with a low quality factor.