Abstract: In this paper, a new method is proposed for digital camera identification from its color images using image sensor noise. Currently the proposed camera identification methods use wavelet-based denoising filter to extract the sensor noise feature. However, the wavelet methods may smooth the edged while denoising and this will lead to low accuracy for those images including highly textured regions. In order to overcome some inherent limitations of wavelet transform, we use curvelet-based denoising filter to obtain the camera fingerprint. Experimental results show that this method provides higher accuracy than other methods on the condition of using a few color images to compute reference pattern, especially for those color images including highly textured regions.

@inproceedings{Zhang:2009ac,
  url          = {http://www.computer.org/portal/web/csdl/doi/10.1109/ICASSP.2009.4959852},
  booktitle    = {Proceedings of the 2009 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2009)},
  author       = {Chi Zhang and Hongbin Zhang},
  year         = {2009},
  title        = {Digital camera identification based on curvelet transform},
  pages        = {1389--1392},
}