Title :
Detection of digital processing of images through a realistic model of CCD noise
Author :
Maillard, J.-B. ; Lévesque, D. ; Deschênes, F.
Author_Institution :
Univ. de Sherbrooke, Sherbrooke, QC
Abstract :
In this paper, we propose a method for detecting digital processing of video such as compositing. Our method is based on a realistic model of charge coupled device (CCD) sensor noise. It allows the identification of an object that has been composited with an image. This is accomplished through the comparison of local characteristics of noise. These characteristics should be constant for an image captured by one camera. However, when an element captured with a different camera is composited with an image, the local noise characteristics of the foreign element can help us identify it. Experimental results validate our model.
Keywords :
CCD image sensors; video signal processing; CCD noise; charge coupled device sensor noise; digital image processing detection; digital video processing detection; Cameras; Charge coupled devices; Charge-coupled image sensors; Digital images; Electrons; Law; Legal factors; Lighting; Sensor arrays; Sensor phenomena and characterization;
Conference_Titel :
Pattern Recognition, 2008. ICPR 2008. 19th International Conference on
Conference_Location :
Tampa, FL
Print_ISBN :
978-1-4244-2174-9
Electronic_ISBN :
1051-4651
DOI :
10.1109/ICPR.2008.4761394