Authors:
Jaroslaw Bernacki
;
Marek Klonowski
and
Piotr Syga
Affiliation:
Wroclaw University of Science and Technology, Poland
Keyword(s):
Hardwaremetry, Camera Recognition, PRNU, Privacy, Histogram Analysis.
Related
Ontology
Subjects/Areas/Topics:
Data and Application Security and Privacy
;
Data Protection
;
Digital Forensics
;
Information and Systems Security
;
Information Assurance
;
Information Hiding
;
Privacy
Abstract:
In this paper we consider the issue of tracing digital cameras by analyzing pictures they produced. Clearly,
the possibility of establishing if a picture was taken by a given camera or even if two pictures come from
cameras of the same model can expose users’ privacy to a serious threat. In the paper, at first, we propose
a simple and ultra-fast algorithm for identification of the brand of a digital camera and compare the results
with the state-of-the-art algorithm by Lukás et al.’s. Experimental results show that at the cost of a moderate
decrease of accuracy, our method is significantly faster, thus can be used for an analysis of large batches
of pictures or as a preprocessing for more exact. In the second part, we propose a method for limiting the
possibility of tracing digital cameras. Our method is a transformation of the picture that can be classified
as a type of standardization. We prove that it prevents all methods of tracing cameras based on analysis of
histograms. Moreover
, in an extensive experimental evaluation we demonstrate that the transformed pictures
are very similar to the original images under both visual and objective numerical measure inspection.
(More)