Single View Metrology And Exposing Digital Forgeries Using Shadows Geometry

With the highly development of image editing softwares and the popularization of digital image device, it is becoming easier and easier to generate highly photorealistic images which erode our trust on the integrity of the visual imagery. Doctored photographs are appearing with a growing frequency in magazines, critical photo evidence in courtroom, insurance claims, science discovery and security surveillance applications. Hence we need to research on more special methods based on computer technology—digital image authenticate technology to guaranteed that the public trust order, crime crackdown, judicial justness and Information Integrity. The past decade has seen considerable advances in authenticate forgeries on single images. Many watermarking-based image authentication schemes have been proposed in recent years. However, a watermarking must be inserted at the time of recording, which would limit the approach to specially equipped digital cameras having no capabilities to add a watermarking when the image capture. On the other hand, the watermarking will be destroyed during image processing which will affect the identification. Therefore, how to validate the image using the characters we extract from the picture in the absence of any protect or prior information is regarded as the new direction. So recent years, the using of the Computer Vision technology is considered as the new direction which attract a lot of attention and received wide applications in image processing, image recognition, scene analysis, image understanding which in order to image preprocessing or extracted statistical characteristics and structure information from the image.This paper introduce some single view metrology such as imaging geometry and homography which based on geometry information to gain the 3D object's height or the space point's world coordinates. Propose a new method which using the shadow parallel property and shadow geometry relationship to gain 3D object's height or the isolated space point's world coordinates. Finally, we exposing fakes by comparing the metric information or match the illuminate consistency and shadow geometry relationship.The input of the proposed methods is a single image to be authenticated. In the first contrastive experiments we implement the measurement methods of 3D object's height we introduced in this paper and the new proposed, the result indicate that the proposed method is easier and simular accuracy. In the second set of contrastive experiments we compare the introduced measurement methods of space point's world coordinates with the new proposed, the result indicate that the proposed is feasible when the point is isolated and simular accuracy. Finally, we validate fakes by the metric information and match the illuminate consistency and shadow geometry relationship.