The Library
Digital image authentication
Tools
Li, Yue (2009) Digital image authentication. PhD thesis, University of Warwick.
Research output not available from this repository.
Request-a-Copy directly from author or use local Library Get it For Me service.
Official URL: http://webcat.warwick.ac.uk/record=b2334464~S15
Abstract
Due to the fast growth of digital technology, daily activities can be easily captured and
saved in digital images, and then transmitted via the Internet. Therefore, the huge amount
of digital images need to be protected through authentication techniques. This thesis is
concerned with digital image authentication (DIA) techniques used for anti-falsification,
copyright protection and access control. DIA techniques can be generally categorised
into data-hiding-based digital image authentication (DHBDIA) techniques and
non-data-hiding-based digital image authentication (NDHBDIA) techniques, which are
both studied in this thesis.
DHBDIA techniques, which are also called watermarking schemes, realise image
authentication by imperceptibly embedding a piece of authentication information into
images. In this thesis, three watermarking schemes are proposed to serve different
purposes. To improve the performance of current clustering-based watermarking
schemes, an expandable progressive exponential clustering (EPEC)-based watermarking
scheme is proposed, which embeds secret data in the host image by progressively and
exponentially clustering the colour intensities in the spatial domain, expanding the
cluster table under predetermined constraints and substituting the intensities in the same
clusters. This scheme strikes a better balance between embedding capacity and
embedding distortion than traditional clustering-based watermarking schemes. To
enhance the security of current quantisation-based watermarking schemes, an adaptive
quantisation index modulation (AQIM)-based watermarking scheme is proposed, which
adaptively modifies the quantised values in the embedding phase, consequently
preserving the histogram of the image and closes the security gap. For the image
authentication, a repetitive index modulation (RIM)-based watermarking scheme is
proposed to set up a non-deterministic dependence among pixels. This scheme achieves
higher security and lower distortion than traditional LSB-based watermarking schemes.
Meanwhile, to study the applicability of DHBDIA techniques, a multifunction digital
image authentication system for medical image management on PACSs is proposed to
perform three different security assurance functions, namely role-based access control,
patient’s medical information integration with mammograms and forgery detection.
In the study of the NDHBDIA techniques, a colour-decoupled photo response
non-uniformity (CD-PRNU) is devised to serve as a camera fingerprint for both source
camera identification and content integrity verification. In the image acquisition process,
colour filtering and interpolation are two important operations that introduce large
amounts of noise, which are not included in the traditional PRNU definition. The
proposed CD-PRNU extends the traditional photo response non-uniformity (PRNU)
extraction model by including such types of noise. As a result, the CD-PRNU is more
accurate and effective in camera identification and content integrity verification. The
main feature of the proposed CD-PRNU extraction algorithm is that it can separate
physical colour components and artificial colour components without the prior
knowledge about the Colour Filter Array (CFA) and then extract the PRNU from each
individual colour channel, thus preventing interpolation noise of artificial components
from diffusing into the sensor pattern noise of the physical colour components.
Item Type: | Thesis (PhD) | ||||
---|---|---|---|---|---|
Subjects: | T Technology > TA Engineering (General). Civil engineering (General) | ||||
Library of Congress Subject Headings (LCSH): | Image processing -- Digital techniques, Digital images -- Watermarking, Copyright -- Computer files | ||||
Official Date: | November 2009 | ||||
Dates: |
|
||||
Institution: | University of Warwick | ||||
Theses Department: | Department of Computer Science | ||||
Thesis Type: | PhD | ||||
Publication Status: | Unpublished | ||||
Supervisor(s)/Advisor: | Li, Chang-Tsun ; Bhalerao, Abhir | ||||
Extent: | 213 leaves : ill., charts | ||||
Language: | eng |
Request changes or add full text files to a record
Repository staff actions (login required)
View Item |