You are on page 1of 7

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882

Volume 5, Issue 4, April 2016

Duplicate Object Detection for Digital Image Authentication


Kusam Sharma1, Pawanesh Abrol2
1, 2

Department of Computer Science & IT, University of Jammu, Jammu

ABSTRACT
Digital Image Authentication has become a very
important research area. Digital images can be altered
easily using freely available image editing softwares.
Many a times, it becomes impossible to ascertain the
authentic image from the tampered image. These
tampered images can be created by using graphical
softwares, or can be manipulated simply by changing the
content and context of the image. This creates problems
for authentication of the digital image. Tampering a
digital image by duplicating an object or a particular
segment within the same image is one of the most
prominent and easily done forgeries. It is known as
copy-move tampering. It is very difficult to find out the
duplicated objects or segments within the given image.
In this paper, we study the recently developed major
approaches to detect duplicate objects in digital images.
Initially, the process of digital image tampering is
explained. Subsequently, we analyze some of recently
developed algorithms like Discrete Wavelet Transform
(DWT), Singular Value Decomposition (SVD), FourierMellin Transform (FMT), Discrete Cosine Transform
(DCT), Principal Component Analysis (PCA), Scale
Invariant Features Transform (SIFT), Wavelets and Logpolar Mapping, Block Artifact Grid (BAG), Speedup
Robust Features (SURF) etc. for detecting duplicate
object within the same image. Preliminary investigations
show that different algorithms have different domains of
tampering detection and have different merits and
demerits. The decision about the content authenticity is
complex and can be better established by interpreting the
results obtained by applying a set of these methods.
Keywords - Copy-Move Tampering, Digital Image
Authentication, Digital Image Tampering, Duplicate
Region Detection

I.

INTRODUCTION

With the advancements in technology, low cost and easy


availability of imaging tools and latest editing softwares
like Adobe Photoshop, GIMP, ProShow Gold, Xn View,

Coral Draw, Free Hand etc., its not difficult now a days
to manipulate digital images to hide or create misleading
images. This creates problem for authenticating the
digital image. There are no established methodologies to
verify the authenticity and integrity of digital images in
automatic manner. So, Digital Image Authentication has
become a very important research area. Tampering a
digital image by duplicating an object or a particular
segment within the same image is one of the most
prominent and easily done forgeries. All of these
tampering can be categorized into three major groups,
based on the process involved in creating the fake digital
image, which are Image Retouching, Image Splicing,
and Copy-Move Attack.
Image retouching does not significantly change an
image, but instead, enhances or reduces certain feature
of an image. It is considered to be the less harmful kind
of digital image tampering. This technique is popular
among magazine photo editors. This technique is
employed to enhance certain features of an image so that
it becomes more attractive [1]. Image splicing, as its
name implies, is a simple process of cropping and
pasting regions from the same or different images to
form a composite image without post-processing such as
edge smoothing. Image splicing is one of the simple and
commonly used image tampering schemes. Since
splicing is often used for image tampering as an initial
step, and splicing itself, with modern image processing
techniques, can often hardly be caught by the human
visual system [2]. This technique is more aggressive
than image retouching.
Whereas copy-move attack is one of the most common
and easiest techniques for creating digital tampering. It
is more or less similar to image splicing in view of the
fact that both techniques modify certain image region of
a base image, with another image or an object. However,
instead of having an external object or an image as the
source, copy-move attack uses portion of the original
base image as its source. In other words, the source and
the destination of the modified image originated from
the same image. In a copy-move attack, parts of the

www.ijsret.org

270

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 5, Issue 4, April 2016

original image is copied, moved to a desired location,


and pasted as another constituent part of the same image
changing the content and context of the same image.
This is usually done in order to conceal certain details or
to duplicate certain aspects of an image. Blurring is
usually applied along the border of the modified region
to reduce the effect of irregularities between the original
and pasted region. It is very difficult to find out the
duplicated objects or segments within the given image.
A lot of research is being done in this domain.
The images can be tampered during the transmission or
it can be tampered later on. So in order to verify the
authenticity and integrity of digital images, various
researchers have proposed different tamper detection
algorithms based on active and passive approach. In
active approach, the digital image requires some preprocessing such as watermark embedding or digital
signature generation at the time of creating the image,
which would limit their application in practice. Whereas
passive approach can be used to find the authentication
of the images without digital signature or watermark
embedded in them. However, unlike techniques in active
approach, techniques in passive approach cannot
guarantee that an image that passed the test is authentic
one or not. Copy-move tampering detection method is
one such most popular passive approach for detecting
duplicated object in a digital image [1].
Techniques in passive approach for detecting image
tampering can further be divided into two main
categories, one that is based on the pixel value of the
image called Statistical Method and one that tries to
detect inconsistencies in the image itself based on visual
cues called Visual Method. Visual method is the easiest
because sometimes it can be performed without the need
for special hardware. One way of detecting traces of
image tampering by visual method is to compare lighting
information from different parts of an image. If the
image is using sun as the primary source of light,
inconsistencies in lighting suggests that possible
tampering has been made because it is nearly impossible
to have two parts of an image with inconsistent lighting
information like different direction of source light [1].
The level of details in tampered image varies depending
on various factors which include the competency of the
forgerer with computer graphics editing software, the
time that he or she has, and the quality of the original

image. Sometimes, the quality of a faked image is so


convincing that it is impossible to use visual method to
detect traces of image forgery. In these cases, techniques
which are based on statistical method can be used to
analyze these kinds of images. The statistical method is
more robust than visual method because it does not
depend on the visual information of an image, but
instead analyses an image based on the values of each of
the pixels in an image.
Detecting duplicated regions in digital images is quite
challenging because of a variety of reasons like type of
image, its size, compression, softwares used to duplicate
the objects and regions, color, grey levels different
image formats, realistic looks as the segments are the
part of the same image, threshold of neighborhood for
verification etc. Further, the detection can become more
difficult if the duplicated object is further subject to
rotation, scaling, blurring, brightness adjustment,
contrast enhancement etc or any cascade combinations
of them.
There are different duplicate region detection methods
that are being used for copy-move tampering. These
methods identify the duplicated regions using different
methods those work efficiently without and with scaling
and rotation are Discrete Cosine Transform (DCT),
Principal Component Analysis (PCA), Discrete Wavelet
Transform (DWT), Singular Value Decomposition
(SVD), Wavelets and Log-polar Mapping and Scale
Invariant Features Transform (SIFT), Block Artifact
Grid (BAG), Fourier-Mellin Transform (FMT), Speedup
Robust Features (SURF) etc respectively.
In this research work, we have studied some of the
recently developed major techniques to detect duplicate
objects in digital images. Initially, the process of digital
image tampering is explained followed by digital image
tampering detection approaches. In the next section, we
discuss some of the recently developed algorithms which
have been presented by different researchers for
detection of duplicate region in digital images. Under
Results & Discussion, we investigate and comparatively
analyze some of recently developed algorithms for
detecting duplicate object within the same image as
mentioned above. Preliminary investigations show that
different algorithms have different domains of tampering
detection and have different merits and demerits. The
decision about the content authenticity is complex and

www.ijsret.org

271

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 5, Issue 4, April 2016

can be better established by interpreting the results


obtained by applying a set of these methods. Finally, we
present the conclusion and the future directions in which
we are working.

II.

RECENT WORK

Digital images are prone to modifications. With the


availability of powerful and easy to use computer
graphics editing software to end users makes the job of
manipulating image easier than ever. Anyone having the
basic knowledge of digital image and the tools in a
computer graphics editing software will be able to
modify an image with ease. Faked image can be used to
misrepresent something, to boost a negative or positive
image of someone. This is where Digital Image
Forensics comes in. Digital Image Forensics deals with
the study of determining the authenticity and
trustworthiness of digital images [1]. Basically there are
three main functions of digital image forensics; Image
source identification, Detection of computer generated
image, and Image tampering detection.
Image source identification focuses on identifying the
primary source digital devices like cameras, mobile
phones, cam carders, scanners etc, using the media
produced by them. Whereas, computer generated images
are to be detected and differentiated from real noncomputer generated images as the level of realism in
these images so high that it cause confusion to the
viewers while tampering detection attempts to discover
the evidence of tampering by assessing the authenticity
of the digital media. Farid et al. classified tampered
images into six categories viz., composite images,
marked images, retouched images, enhanced images,
computer graphics and painted images [3].
For the detection of such tampered images, various
researchers have given different tamper detection
algorithms which works in different domains. One such
approach is based on Discrete Wavelet Transform
(DWT) for detection purpose [4]. Its a blind forensics
approach for detecting Copy-Move tampering which
works by first applying DWT i.e. Discrete Wavelet
Transform to the input image to yield a reduced
dimension representation. Then the compressed image is
divided into overlapping blocks. These blocks are then
sorted and duplicated blocks are identified using Phase

Correlation as similarity criterion. Due to DWT usage,


detection is first carried out on lowest level image
representation. This approach drastically reduces the
time needed for the detection process. Since exhaustive
comparisons of blocks have been applied only on the
image in the lowest resolution level, the approach has
significantly improved the time and accuracy of
detection compared to the past techniques of CopyMove forgery. This algorithm is robust to common postprocessing operations like noise and JPEG quality level
changes. But, the main drawback of this approach is that
it cannot detect the duplicated regions with rotation
through angles and scaled regions.
Another blind forensics approach is based on Discrete
Wavelet Transform (DWT) & Singular Value
decomposition (SVD) to detect the specific artifact [5].
In this, the image is first reduced in dimension by DWT
and then the SVD is applied to the fixed-sized
overlapping blocks of low frequency wavelet portion.
The singular value vectors are then lexicographically
sorted and duplicated image blocks will be close in the
sorted list, and therefore will be compared during the
detection steps. This approach can not only decrease
computational complexity, but also localize the
duplicated regions accurately even when the image is
highly compressed or edge processed. Another improved
detection algorithm based on Discrete Wavelet
Transform (DWT) and Principal Component Analysis
Eigen Value Decomposition (PCA-EVD) has been
presented by Zimba in [6].
A new technique based on Fourier-Mellin transform
(FMT) has come into existence [7]. In this technique, the
fourier-mellin transform representation of each block is
obtained and re-sampled the resulting magnitude values
into log-polar coordinates. Then, a vector representation
is obtained by projecting log-polar values onto 1-D
which are used as features. These features would not
only robust to lossy JPEG compression, blurring, or
noise addition, but also known to be scaling and
translation invariant. Furthermore, the detection time is
reduced by using counting bloom filters, instead of
lexicographic sorting. These counting bloom filters
essentially compare the hashes of features instead of
features themselves.

www.ijsret.org

272

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 5, Issue 4, April 2016

Realization of Counting Bloom Filters and


Tampering Decision:
To improve the efficiency of detection, counting bloom
filters are used, which essentially compares the hashes of
features as opposed to features themselves [7]. This is
realized by the equations (1) and (2) as following:
Form an array K with k elements which are all
zero initially.
Hash the feature vector fi of each block such
that each hash value will indicate an index
number in the array K.
If the feature vectors of two blocks are identical
they would give the same hash value yielding
same index value, increment the value of the
corresponding element in K. That is,
h = hash(fi)
(1)
K(h) = K(h) + 1

(2)

Any element of array K that is higher than 2


indicates duplicated block pairs.
Popescu et al. proposed a technique based on Principal
Component Analysis (PCA). Its another efficient
technique which can automatically detect and localize
duplicated regions in an image [8]. It works by first
applying a Principal Component Analysis (PCA) on
small fixed-size image blocks to yield a reduced
dimension
DCT
block
representation.
This
representation is robust to minor variations in the image
due to additive noise or lossy compression. In this, each
block is represented as 16x16 and the coefficients in
each block are vectorized and inserted in a matrix and
the corresponding covariance matrix is constructed using
equation (3):
-

(3)

The matrix so constructed stores floating numbers. By


finding the eigenvectors of the covariance matrix, a new
linear basis is obtained. The duplicated regions are then
detected by lexicographically sorting all of the image
blocks. This technique is effective on plausible forgeries,
and has quantified its sensitivity to JPEG lossy
compression and additive noise.
Fridrich et al suggested a technique based on Discrete
Cosine Transform (DCT) coefficients [9]. In this
method, the possibility for blind tampering detection is
to classify textures that occur in natural images using

statistical measures and find discrepancies in those


statistics between different portions of the image. It
supports two algorithms for detecting Copy-Move
tampering, one that uses an exact match for detection
and other that is based on an approximate match. The
two approaches introduced by the approximate match
algorithm are Exhaustive Search and Autocorrelation
whereas two other approaches introduced are Exact
match algorithm and Robust match algorithm. In this
method, first the image is segmented into overlapping
small blocks followed by feature extraction using
discrete cosine transform (DCT) coefficients. The DCT
coefficients of the small blocks are then
lexicographically sorted to check whether the adjusted
blocks are similar or not. This method is robust to the
retouching operations. Another DCT-SVD based method
used for detection of copy-move forgery based on
overlapping and non-overlapping technique is given by
Zhao in [10]. This method is quite efficient in detecting
duplicated regions even in the presence of Gaussian
blurring, AWGN, JPEG compression.
The problem of detection of digital image forgeries is a
critical issue having no universally applicable solution.
Thus, a set of different image tampering detection tools
can be applied to the image at hand. The decision about
the content authenticity is then reached by interpreting
the results obtained from different approaches. So in
future, all these techniques in conjunction with a
growing body of other forensic tools, is effective in
exposing digital tampering.
Also based on above mentioned approaches, we have
discussed about the general working of a Copy-Move
Digital Image Tampering Detection System given by
Shivakumar et al. in which the inputted image is divided
into overlapping blocks to detect connected blocks that
are copied and moved [11]. This copied region is further
consisted of many overlapping blocks. Since each block
is moved with same amount of shift so the distance
between each duplicated block pair would be the same.
The next phase deals with the extraction of features form
these blocks, which would yield to very similar or same
values for duplicated block. Several authors presented to
use different features to represent the image block using
various approaches like DCT, DWT, PCA, SVD, FMT,
etc. These blocks are vectorized and inserted into a
matrix and the vectors are lexicographically sorted for

www.ijsret.org

273

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 5, Issue 4, April 2016

later detection. The computational time depends upon


factors such as number of blocks, sorting techniques and
the number of features. The vectors corresponding to
blocks of similar content will be close to each other in
the list, so that the identical regions can be easily
detected. Figure 1 shows the block diagram of CopyMove Digital Image Tampering Detection System.

Fig.1: Block diagram of a Copy-Move Digital


Image Tampering Detection System

III.

SUGGESTIONS
WORK

AND

FURTHER

Based on the study, it is evident that there are different


algorithms that help detect the different types of
tampering like Discrete Wavelet Transform (DWT)
which works by first applying DWT i.e. Discrete
Wavelet Transform to the input image to yield a reduced
dimension representation. Then the compressed image is
divided into overlapping blocks. These blocks are then
sorted and duplicated blocks are identified using Phase
Correlation as similarity criterion, Wavelets and Logpolar Mapping works by applying wavelet transform to
detect and perform exhaustive search to identify the
similar blocks in the image by mapping them to logpolar coordinates and using phase correlation as the
similarity criterion [12], Singular Value Decomposition
(SVD) is used to detect the specific artifact. In this, the
image is first reduced in dimension by DWT and then
the SVD is applied to the fixed-sized overlapping blocks
of low frequency wavelet portion. The singular value

vectors are then lexicographically sorted and duplicated


image blocks will be close in the sorted list, and
therefore will be compared during the detection steps,
Fourier-Mellin Transform in which FMT representation
of each block is obtained and re-sampled the resulting
magnitude values into log-polar coordinates, Discrete
Cosine Transform (DCT) in which the possibility for
blind tampering detection is to classify textures that
occur in natural images using statistical measures and
find discrepancies in those statistics between different
portions of the image, Principal Component Analysis
works by first applying PCA on small fixed-size image
blocks to yield a reduced dimension DCT block
representation, Scale Invariant Features Transform
(SIFT) in which choice on the number of clusters is to be
made sensitively [13], Block Artifact Grid (BAG) in
which if the copied area is from the other different
image, it cannot be detected by the method [14],
Speedup Robust Features (SURF) etc.
Among all mentioned above, one of the important areas
is DCT (Discrete Cosine Transform). Many factors are
important here like type of image, degree of tampering,
size of image, extent of tampering, compression,
softwares used to duplicate the objects and regions,
color, grey levels, different image formats, realistic
looks as the segments are the part of the same image,
threshold of neighborhood for verification etc. Based on
this, we are working on Digital tampering detection and
we are working on to develop a frame work for
ascertaining the extent of tampering in the given digital
image.

IV.

RESULTS AND DISCUSSION

The practice of tampering photographs is probably as old


as the art of photography itself. Digital photography and
powerful image editing softwares like Adobe Photoshop,
Xnview, ProShow Gold, made it very easy today to
create believable tamperings of digital pictures even for
a non-specialist. As digital photography continues to
replace its analog counterpart, the need for reliable
detection of digitally doctored images is quickly
increasing.
Recently, several different methods for detecting digital
tamperings were proposed like Discrete Cosine
Transform (DCT), Principal Component Analysis

www.ijsret.org

274

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 5, Issue 4, April 2016

(PCA), Discrete Wavelet Transform (DWT), Singular


Value Decomposition (SVD), Wavelets and Log-polar
Mapping and Scale Invariant Features Transform
(SIFT), Block Artifact Grid (BAG), Fourier-Mellin
Transform (FMT), Speedup Robust Features (SURF) etc
respectively.
Obviously, the problem of detection of digital forgeries
is a complex one with no universally applicable solution.
Thus, a set of different tools can be all applied to the
image at hand. The decision about the content
authenticity is then reached by interpreting the results
obtained from different approaches. This accumulative
evidence may provide a convincing enough argument
that each individual method cannot. So in future, all
these techniques in conjunction with a growing body of
other forensic tools, is effective in exposing digital
tampering.
The comparative analysis of some of the selected
recently developed above mentioned algorithms on the
basis of their merits-demerits, domain, types of inputoutput etc. has been presented in the form of table, Table
1.

V.

tampering detection for detecting duplicate object within


the same image using different algorithms. Our current
study is based on studying different copy-move
detection methods having different merits and demerits.
However, there are many research issues which are
required to be addressed while detecting duplicate object
within the same image using different algorithms.

VI.

FUTURE SCOPE

Our further work consists of building a frame work for


duplicate object detection for authenticating digital
images using non-overlapping block-based tampering
detection technique. This work can further be carried out
to analyze other recent algorithms related to Duplicate
Region Detection used for Copy-Move tampering like
Scale Invariant Features Transform (SIFT), Block
Artifact Grid (BAG), Speedup Robust Features (SURF)
etc. The algorithm may be optimized for faster detection.
Further, the algorithms related to the detection of other
types of image artifacts, such as spliced images,
photorealistic computer graphics etc. are to be studied
and analyzed.

REFERENCES

CONCLUSION

The given research work studies the recently developed


major approaches to detect duplicate objects in digital
images. The copy-move tampering detection is one of
the emerging problems in the field of digital image
forensics. Many techniques have been proposed to
address this problem. The sophisticated and low-cost
tools of the digital age enable the creation and
manipulation of digital images without leaving any
perceptible traces. As a result, the authenticity of images
cant be taken for granted, especially when it comes to
legal photographic evidence. One of the biggest issues
these techniques had to deal with was, being able to
detect the duplicated image regions without getting
affected by the common image processing operations,
e.g. compression, noise addition, rotation. The other
challenge is computational time, which becomes
important considering the large databases. Initially in
this paper, the process of digital image tampering is
explained. Subsequently, we have studied and analyzed
some of the selective algorithms to know the recent
developments in the field of Copy-move digital image

[1] W. Lu, W. Sun, J. W. Huang, and H. T. Lu, Digital image


forensics using statistical features and neural network
classifier, International Conference on Machine Learning
and Cybernetics, 2008 IEEE, PP 2831-2834.
[2] Y.Q. Shi, C. Chen, and W. Chen, A Natural Image Model
Approach to Splicing Detection, 2007 ACM, pp 2021.
[3] K. Sharma, and P. Abrol, Non-Overlapping Block-based
Copy-Paste Forgery Detection Model, International Journal of
Computer Applications, 133(3), 2016, 17-24.
[4] K. Sharma, and P. Abrol, Transformation Based
Parametric Analysis for Copy-Paste Tampering Detection,
International Journal of Scientific and Technical
Advancements, 2(1), 2016, 247-254.
[5] M. Ishrat, K. Pawar, and P. Abrol, Detection and response
to external stimuli, Journal of Scientific and Technical
Advancements, 1(3), 2015, 97-99.
[6] H. Farid, Creating and Detecting Doctored and Virtual
Images: Implications to the Child Pornography Prevention
Act, Technical Report TR2004-518, Department of
Computer Science, Dartmouth College, 2004.
[7] S. Khan, and A. Kulkarni, An Efficient Method for
Detection of Copy-Move Forgery Using Discrete Wavelet
Transform, International Journal on Computer Science and
Engineering, 2(5), 2010, 1801-1806.
[8] G. Li, Q. Wu, D. Tu, and S. Sun, "A Sorted Neighborhood
Approach for Detecting Duplicated Regions in Image

www.ijsret.org

275

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 5, Issue 4, April 2016
Forgeries based on DWT and SVD, International Conference
on Multimedia and Expo, 2007 IEEE, pp. 1750-1753.
[9] D. Sharma, and P. Abrol, Investigating the Extent of Noise
in Digital Images using SVD, International Journal of
Software and Web Sciences, 4(1), 2013, 6-14.
[10] M. Zimba and S. Xingming, DWT-PCA (EVD) based
copy-move image forgery detection, International Journal of
Digital Content Technology and its Applications, 5(1), 2011,
pp. 251-258.
[11] S. Bayram, H. T. Sencar, and N. Memon, An efficient
and robust method for detecting copymove forgery,
International Conference on Acoustics, Speech and Signal
Processing, 2009 IEEE, pp. 10531056.
[12] A. C. Popescu and H. Farid, Exposing Digital Forgeries
by Detecting Duplicated Image Regions, Technical Report
TR2004-515, Department of Computer Science, Dartmouth
College, 2004.
[13] J. Fridrich, D. Soukal, and J. Luk, Detection of CopyMove Forgery in Digital Images, Proceedings of the Digital
Forensic Research Workshop, 2003.

[14] J. Zhao and J. Guo, Passive forensics for copy-move


image forgery using a method based on DCT and SVD,
Elsevier, Forensic Science International, 233(1-3), 2013, pp.
158-166.
[15] B.L. Shivakumar, and S.S. Baboo, Detecting Copy-Move
Forgery in Digital Images: A Survey and Analysis of Current
Methods, Global Journal of Computer Science and
Technology, 10(7), 2010, pp 61-65.
[16] A.N. Myna, M.G. Venkateshmurthy, and C.G. Patil,
Detection of region duplication forgery in digital images
using wavelets and log-polar mapping, Proceedings of the
International Conference on Computational Intelligence and
Multimedia Applications, 2007 IEEE, pp. 371377.
[17] H. Huang, W. Guo, and Y. Zhang, Detection of CopyMove Forgery in Digital Images Using SIFT Algorithm,
Pacific-Asia Workshop on Computational Intelligence and
Industrial Application, 2008 IEEE, pp 272-276.
[18] W. Li, N. Yu and Y. Yuan, Doctored JPEG Image
Detection, International Conference on Multimedia and
Expo, 2008 IEEE, pp 253-256.

Table 1: Comparison Of Different Duplication Region Detection Algorithms


Duplication Region
Detection Algorithms based
on /Methods Used

1. Discrete Cosine
Transform (DCT)
Coefficients

2. Discrete Wavelet
Transform (DWT)

3. Discrete Wavelet
Transform (DWT) &
Singular Value
Decomposition (SVD)

4. Principal Component
Analysis (PCA)

5. Fourier-Mellin Transform
(FMT)

Merits

Demerits

a) The advantage of DCT is that the signal energy would be concentrated


on the first few coefficients, while most other coefficients are negligibly
small.
b) Therefore, the changes in high frequencies, which would occur due to
the operations such as noise addition, compression and retouching should
not affect these first coefficients.

a) This technique is robust to the retouching


operations.
b) Exhaustive search technique used for
detection is quite computationally expensive.

a) This approach drastically reduces the time needed for the detection
process.
b) Since exhaustive comparisons of blocks have been applied only on the
image in the lowest resolution level, the approach has significantly
improved the time and accuracy of detection compared to the past
techniques of Copy-Move forgery.
a) This algorithm decreases computational complexity,
b) It also localizes the duplicated regions accurately even when the image
was highly JPEG compressed or edge processed to a certain extent.
c) This approach can further reduce image scale and feature dimension
while retaining its efficiency.
d) In comparison to PCA and DCT approaches, this approach can better
improve detection efficiency.
a) This technique is effective on plausible forgeries, and has quantified
its sensitivity to JPEG lossy compression and additive noise.
b) This method is robust to compression up to JPEG quality level 50 and
the time complexity of sorting is O(32x k lg k) time.
c) The detection is possible even in the presence of significant amounts
of corrupting noise.
a) It is robust to compression up to JPEG quality 20, rotation with 10 o and
scaling by 10%.
b) The proposed algorithm can detect duplicated region in the images
very accurately, even when the copied region has undergone severe image
manipulations like lossy compression, scaling and rotation.

www.ijsret.org

a) The performance relies on the location of


copy-move regions.
b) This algorithm cannot detect the duplicated
regions with rotation through angles and scaled
regions.

a) The method fails if quality factors of JPEG


compression is lower than 50.
b) The computation of SVD takes lot of time
and it is computationally complex.

a) May fail to detect considerable large changes.


b) Little doubt is there that counter-measures
will be created to foil this technique

a) The algorithms works in case of only slight


rotation.

276

You might also like