Image Authentication by Detecting Traces of Demosaicing 


Can you spot the tampered regions in the images on the left? Our algorithm localizes the tampered regions by finding image portions that do not exhibit traces of demosaicing (and are therefore likely to be forged). The middle column shows our analysis signal, and the right column shows the suspicious image regions. The original images are shown on the right. A careful look shows our algorithm accurately localized the tampered image regions. The algorithm is also used to distinguish between computer generated images and photographic images.


With increasing technical advances, computer graphics are becoming more photorealistic. Therefore, it is important to develop methods for distinguishing between actual photographs from digital cameras and computer generated images. We describe a novel approach to this problem. Rather than focusing on the statistical differences between the image textures, we recognize that images from digital cameras contain traces of resampling as a result of using a color filter array with demosaicing algorithms. The in-camera processing (rather than the image content) distinguishes the digital camera photographs from computer graphics. Our results show high reliability on a standard test set of JPEG compressed images from consumer digital cameras. Further, we show the application of these ideas for accurately localizing forged regions within digital camera images.

Andrew Gallagher
Tsuhan Chen

A. Gallagher, T. Chen, "Image Authentication by Detecting Traces of Demosaicing," IEEE Conference on Computer Vision and Pattern Recognition, Workitorial on Vision of the Unseen 2008 (to Appear).



author = {A. Gallagher and T. Chen},

title = {Image Authentication by Detecting Traces of Demosaicing},

booktitle = {Proc. CVPR WVU Workshop},

year = {2008},


A Small Tampered Image Database 
As part of this work, three original images were tampered with to produce three images with forged regions.
Original Images: Kids, Sheep, Street
Tampered Images: KidsM, SheepM, StreetM
Note that the Sheep image is publically available from a generous Flickr member (Mark Wales). The Street 
image is from Columbia's ADVENT image set (T.-T. Ng, S.-F. Chang, J. Hsu, and M. Pepeljugoski.
Columbia photographic images and photorealistic computer graphics dataset. Technical report, Columbia
University, 2005.).

Please cite both my paper and the Columbia paper if you use these images.