2018 International Conference on Engineering Technologies and their Applications (ICETA), Islamic University – ALNajaf - IRAQ

Comprehensive Display of Digital Image Copy-Move

Forensics Techniques
1
Jumana Waleed , 2 Dhahir Abdulhade Abdullah ,3 Mokhles Hussein Khudhur
1, 2, 3
Department of Computer Science, College of Science, University of Diyala, Iraq
Email : 1 jumanawaleed@sciences.uodiyala.edu.iq ,2 dhahair@sciences.uodiyala.edu.iq
Email : 3 mukhudhur@gmail.com

Abstract— Digital Images considered as a natural and effective signal. There are different operations which happen through
communication media to the human, owing to the swiftly and the forgery are; cropping, blurring, adding noise, rotation, scaling,
easy manner to comprehend their content. After the appearing of compression, down sampling, resizing, retouching, and etcetera.
the high-quality digital camera, personal computers, and The revolution of digital information and matters related to the
sophisticated photo editing software, the manipulation becomes
security of multimedia has created different approaches in the
more common. One of the most popular tools to manipulation is a
copy-move forgery (CMF) that precedes via copying portion of the field of tampering detection and IF. [2]. The significance of
digital image and paste it into another position in the same image appearing different approaches of forgery detection is, when a
to conceal the object behind. In this paper, a comprehensive situation between two parties, taking any decision depending on
expansion of digital image copy-move detection techniques is the given forged images without having the original images is
presented to display the appropriate techniques which are extremely difficult and will lead to disastrous results. Therefore,
presently used. Many algorithms have been developed, algorithms it is tricky to detect these manipulations. As a result, the
based on DCT, algorithms using invariant image moments, authenticity and integrity of images are lost. The alteration of
algorithms using texture and intensity descriptors, algorithms digital images can be utilized in several malicious purposes
using invariant key points, algorithms based on mutual
such as for hiding some significant traces of an image or to
information, and algorithms based on SVD to determine the
existence of digital image forgery. transmit incorrect information. And to identify the integrity of
these images, there is a necessity for detecting if received
images are forged or not [3].
Keywords— Copy-Move Forgery (CMF), Dependent Forgery, Recently, different approaches have been presented for
Image Forensics (IF), Passive-Blind Approaches, Tampering tracing the digital image forgery. Generally, these approaches
Traces.
are categorized into active and passive-blind. Active
I. INTRODUCTION approaches are classified into the data hiding such as
watermarking [4], [5], [6] and digital signature approaches. In
Nowadays, the image is the most popular manners of
contrast to active approaches, passive-blind approaches work
communication, the image can easily, correctly, and quickly be
without using any techniques of protection and with the absence
carrying any idea between the recipients. The vast domain of
of any previous information concerning the image. For
applications leads the image to be most affected by fraud and
detecting the tampering traces, the blind approaches utilize the
tamper. The swift spread of cheap and simple to use devices
function of image and the fact that the forgery can leave
which qualify the visual data acquisition makes approximately
specified detectable modifications to the image such as
everyone able to record, store, and share many digital images.
statistical traces. This paper focuses on passive-blind
The large availability of image editing software tools makes
approaches especially CMF detection, regarded as a new trend.
extremely simple to alter the content of the images. Therefore,
Different algorithms based CMF are used to detect the digital
there is no confidence that anything appears in a photo is a real
image forgery or image manipulation which had a potential
representation of what truly occurred. The photography value
optimization to provide the accurate decisions without using
should be carefully evaluated as events record. This necessity
any previous knowledge or information of the original image
comes from a various range of applications; The most
as algorithms based on DCT, algorithms using invariant image
significant one is the scenario of forensics, in which the
moments, algorithms using texture and intensity descriptors,
reliability of the image should be confirmed before utilizing it
algorithms using invariant key points, algorithms based on
as a prospective evidence.
mutual information, and algorithms based on SVD to determine
The image forensics (IF) is the science addressing the
the existence of digital image forgery. The paper is structured
identification, validation, analyzing, and interpretation of the
as follows; this section presents a brief introduction to the image
digital images as a prospective evidence [1]. IF aims to
forgery detection approaches and the second section
understand if the given image is a combination of various shots.
summarizes these approaches and focused on passive-blind
Generating a forgery commonly needs some steps of processing.
approaches. Section three shows the recently used copy-move
Permanently, these steps remain some statistical traces in the
978-1-5386-7858-9/18/$31.00 ©2018 IEEE

155
2018 International Conference on Engineering Technologies and their Applications (ICETA), Islamic University – ALNajaf - IRAQ

forensic based algorithms. The performance is analyzing in signature in a digital image; also, the wasting of processing time
section four; And finally, the conclusions are drawn in the last to examine the authenticity at the receiver side [10]. Passive-
section. blind approaches are further categorized into dependent and
independent forgery approaches [8]. The forgery-dependent
II. IMAGE FORENSICS (IF) detection approaches are developed for detecting specific types
The digital forensics domain is developing considerably to of forgery like splicing and copy-move that are based on the
resist the IF problems in different fields such as sports, medical kind of forgeries accomplished on an image. Whilst the forgery-
images, legal services, and intelligence. These digital images independent approaches detect forgery without reliance on the
can be given as proof for the court of law. In such cases, it type of forgeries but depending on tampering traces left through
becomes extremely significant for proving the originality of the operation of resampling and lighting inconsistencies [11].
digital images. IF plays a dynamic role in these cases by
examining authenticity and integrity of digital images [7]. For III. COPY-MOVE FORENSIC BASED ALGORITHMS
proving the authenticity of digital images, different approaches Copy-Move is the most common and popular photographs
have been presented which are generally divided into active and forgery approach due to the ease that can be achieved through
passive approaches; Fig. 1 illustrated the IF techniques. the use of it [7]. It includes the copying of some regions on the
image and moving them to other regions on the image [12] [13]
[14] [15] [16] [17]. Because the copied regions belong to the
same image, so, the dynamic area and color stay consistent with
the remainder of the image [18]. In this section, we try to give
a guide of the researches that appeared in this field by focusing
on the recently used copy-move forensic based algorithms;
showing and comparing the advantage and disadvantage of each
one.
The general algorithm used in the CMF detection should
come as follows:
1. Convert the forged image with size (𝑀 × 𝑁) into a
grayscale.
2. Divided the image into overlapping blocks with size
(𝑏 × 𝑏) to find 𝑂 = (𝑀 − 𝑏 + 1) × (𝑁 − 𝑏 + 1) of
Fig. 1 Digital Image Forensic Techniques. overlapping blocks; And these blocks are rearranged
into a row vector Ri with size (𝑂 × 𝐾); 𝑤ℎ𝑒𝑟𝑒 𝐾 = 𝑏 2 .
A. ACTIVE APPROACHES After that, for each block, the features are extracted
With the active approaches, the authentication process needs and stored as rows depending on the used algorithm.
prior information concerning the image. It is related to 3. Perform the lexicographical sorting on the rows of the
embedding the data where at the generation time a code is extracted feature matrix. This leads to getting identical
hidden into the image. Proving this code will lead to blocks.
authenticate the originality of the digital image. [8].
Manipulation of the image consists of many processing 4. Apply the matching process. In the conventional
operations like scaling, rotating, blurring, brightness adjusting, matching, the pairs of the feature vector are searched
change in contrast, etc. or any combination of these operations. among the nearest neighbors utilizing a threshold. If
Doctoring image means pasting one part of the image into Rij is the matched pair including the features Ri and Rj,
another part of the image, skillfully without living any trace. where i ≠ j represents indices of the feature; After that,
One important tool for the authenticity of the digital image is the shift vector SV between two matching blocks is
the watermarking and digital signature [9].
computed as:
B. PASSIVE-BLIND APPROACHES 𝑆𝑉𝑖𝑗 (𝑑𝑥 . 𝑑𝑦 ) = (𝑋𝑖 − 𝑋𝑗 . 𝑌𝑖 − 𝑌𝑗 ) (1)
Passive-blind approaches are proving the authenticity of the The counter of shift vectors CO(SV) is taken out and
digital images without the need for prior information, only the incremented by one for each matching pair of blocks
image itself. It assumes that although the manipulation may do with the same shift as in the following equation:
not leave any perceptible traces that although the manipulation
𝐶𝑂(𝑑𝑥 . 𝑑𝑦 ) = 𝐶𝑂(𝑑𝑥 . 𝑑𝑦 ) + 1 (2)
may do not leave any perceptible traces, it is probably to change
the implicit statistics. These inconsistencies can be utilized for Matching pairs of blocks with the same shift are
detecting the forgery. The passive-blind approaches became grouped. Ignore the groups of blocks with a count of
extremely significant to overcome the difficulties of active shift vector below 𝜃. The threshold 𝜃 controls the size
approaches represented in the preceding knowing of images of the little copy-move detectable segment.
content, and the time of processing to hide a watermark or

156
2018 International Conference on Engineering Technologies and their Applications (ICETA), Islamic University – ALNajaf - IRAQ

5. Perform morphological operations for minimizing the region with different size and shape. It is very efficient to
false positives. partition the digital image into fixed sized of overlapping blocks
Fig. 2 illustrates the general structure of the copy-move and do the experiment if the blocks pairs are duplicated. The
forensic based algorithms. main advantages of this algorithm that it used fewer features to
represent each block. This algorithm showed an ability to detect
copy-move forgery in an image that is quite robust to JPEG
compression, blurring or white Gaussian noise distortion. But,
the researchers did not mention to one main disadvantage that
this method could not take a decision with rotated or reflected
image. Also, Nathalie Diane W. et al [22], have used the same
methodology to copy-move forgery detection based on DCT,
except the usage of the Euclidean distance as a similarity
criterion to identify the duplicated blocks; This method showed
that the duplicated regions can be detected even there is further
copy-move forged area in the image, in addition to the case of
JPEG compression, small rotations, noise addition, blur, shift,
and scale.

B. ALGORITHMS BASED ON INVARIANT IMAGE MOMENT

The concept of image moments refers to a certain particular
weighted average for the intensities of the image pixel. These
moments are selected for giving a uniform image interpretation
which assists in the analysis of the shape. It is beneficial for
describing the objects of an image after segmentation or
describing the total image central point, intensity, and
information of objects orientation for detecting scaling, rotation,
and translation. The image moments connect regions in a binary
Fig. 2 The general structure of the copy-move detection. shape for translating scaling, and rotation in a suitable
classification shape and part recognition [23]. The translation
A. ALGORITHMS BASED ON DCT invariance is the main characteristic of the centralized moments.
A lot of persons have utilized the block matching for finding Even scaling or rotating the image, the centralized moments are
the duplicates of the region with various features demonstrating still not changed. It is extremely significant for different
one block of the image [19], [20]. Fredrick et al. [12] firstly, applications to construct these moments with scaling, rotation,
offered a technique by utilizing exhaustive search; After that and translation invariance. [24]. CMF can be traced by
suggested an effective block matching detecting technique computing Hu, Blur invariant, and Zernike moments. Babak
depending on discrete cosine transform (DCT). The main idea and Stanislav [25] utilize 24 blur-invariant moments as features
behind using an algorithm based on DCT is to use its for detecting the CMF even in the case of existence contrast
coefficients as a feature to be compared to find the repeated changes in duplicated regions. Wang et al. [26] utilize the Hus
blocks. Popescu [13] suggested a technique which utilizes the moments as features which are robust for posting region
principal component analysis (PCA) rather than DCT. Owing to duplication image processing consisting of noise contamination,
the attributes of PCA, the features needed to represent the block blurring, lossy compression, and rotation. And the first four-
was decreased to approximately half compared with the features moments are utilized as features to reduce the computational
utilized by Fridrich. Therefore, the technique which uses the complexity. Ryu et al. [27] use the Zernike moments of circular
PCA has a preferable time complexity; But this technique has blocks as features to detect the CMF of the regions in the forged
the low robustness to small rotations of copy-moved regions. images. these features are robust to rotation, blurring, adding
Yanping Huang et al. [21] present the usage of an algorithm noise, and compression. The usage of these features is also
based on improved DCT through dividing the image into fixed- helpful to detect the CMF of the blocks for flat regions. The
overlapping blocks and applying DCT on each block for main disadvantage of the algorithms based on moments is the
representing its features. These features are truncated for high computational complexity.
reducing the demotion and lexicographically sorted for
neighbouring the duplicated blocks in the sorted list. Matching
between blocks is applied to detect duplications. Here the C. ALGORITHMS USING TEXTURE & INTENSITY DESCRIPTORS
researchers suppose that the duplicated regions aren't Texture and intensity exist in natural scenes such as grass,
overlapping. The process of the detection is for determining if tree, cloud, and ground, and the characteristics of the image like
the digital image includes duplicated regions. Because the size coarseness, smoothness, and regularity represent the texture
and shape of regions are unknown, it is computationally contents. So, the texture and intensity can be used as features to
impossible for trying to examine each potential pair of the locate the similarity in the tampered image. In CMF detection,

157
2018 International Conference on Engineering Technologies and their Applications (ICETA), Islamic University – ALNajaf - IRAQ

texture and intensity are computed and distinguished by pattern E. ALGORITHMS BASED ON MUTUAL INFORMATION
and intensity [28]. The idea of mutual information was firstly suggested by
A block-based technique utilizes the texture and intensity as Soleimani and Khosravifard [38] in which the template
features for detecting the CMF. The blocks of the image are matching of an image refers to the dependency between two
partition into several directional sub blocks; After that, these random variables. Mutual information is in its maximum value
sub-blocks are used to obtain the energy features. This when two regions or random variables are dependent. In this
technique works with grayscale and colour images [29], [30]. case, the joint probability matrix is diagonal, and in the case of
Quan and Zhang [31] propose a CMF detection technique in independent regions or variables, the mutual information equals
image blocks with similar texture rather than exploiting the zero. Chakraborty [39] proposed a new technique for detecting
whole image for decreasing the computational complexity of the CMF depending on mutual information searches for
the technique. In this technique, firstly the image segmentation duplicated regions with no need to extract any features, depends
is applied by utilizing local dimension estimation and then on a mathematical way. The major advantages of this technique
matches source and target regions in the image with the same that it is easy and high speed and it is perfectly robust against
texture. In [32], Wang et al. utilize the circular blocks and mean illumination alterations.
value of the image pixel in concentric circles of different radii
as features, and the algorithms based on these features are F. ALGORITHMS BASED ON SVD
proven the robustness against rotation. In [6], Ardizzone et al. Singular value feature vectors (SVD) is served to produce
propose a technique which used statistical texture features like algebraic and geometric invariant and feature vectors. The
standard deviation, mean, kurtosis, and skewness, for detecting algorithm divides the image into overlapping blocks. Then to
the CMF whilst the authors in [33] utilized Gabor filter for each block applies the SVD and obtain reduced rank dimension
extracting robust features. and extract singular value feature vectors and store it in a matrix.
Lexicographically sort and consecutive rows indicate identical
D. ALGORITHMS BASED ON INVARIANT KEY POINTS
blocks. When the similarity between blocks is higher than a
The algorithm based on the key point extracts the feature fixed value, the CMF is detected. The method localizes the copy
point from specific regions of an image without any partition of move tampering, has lower computational complexity and
an image [34]. It is excelling in the computation time and higher noise immunity [40]. In [41], The SVD based technique
memory consumption, also it has robustness against rotation, supplies the spatial position of the image part that has been
scaling, Gaussian noise, JPEG compression, and illumination in forged by comparing the recovered image and the forged image
which a block-based algorithm fails. The size of the extracted that the recipient receives.
feature is comparatively large, while, the key points are
extremely smaller in size than that of image blocks. Hence, the IV. PERFORMANCE ANALYSIS
key point based algorithm is being the perfect selection for CMF All the algorithms based CMF detection techniques
detection in the images of large size. It extracts the feature point discussed in the sub-sections above are used to detect the copy-
by utilizing several approaches such as Scale-Invariant Feature move forgery, and a comparison is given in Table 1 among the
Transform (SIFT), Speeded-Up Robust Feature Extraction most of these algorithms. This comparison is depending on the
(SURF), etcetera. with no need to partition the image. The criteria: the used algorithm, the matching strategy, and the
Feature points are matched with each other by utilizing several performance.
approaches such as Euclidean distance and clustering [35]. A Digital image forensics techniques have developed
Forgery detection technique depending on matching image sufficiently to resist the digital image forgery problem in
SIFT is proposed by X. Pan and S. Lyo [36]. This technique is different areas such as medicine, sports, legal services, and
robust to different transformations like scaling, rotation, and intelligence. There are a large number of works have been
less sensitive to JPEG compression and noise. The algorithm appeared in the areas of detecting the digital image CMF, and
includes several steps; First, identify the key-points and collect this is obvious in the Fig. 3 and Fig. 4 that illustrate the papers
the SIFT feature vector for each key-points; Then, SIFT key- amount which addressed the CMF detection techniques in
points are matched and prune matches to reduce the false Science Direct and IEEE in the last ten years.
matches. The estimation of region transforms based on the
correlation coefficient identify the duplicated regions. A
technique based on SURF in flat regions was proposed by G.
Zhang and H. Wang [37]. This technique firstly detects copy-
move forgery by extracting SURF key-points from the image.
Then, the feature matching and pruning are done. Finally,
estimate region transforms and identify duplicated regions
using correlations adjusted using estimated transforms. The
main advantage of this method that it can detect the region
duplication in the non-flat region and is rotation invariant.

158
2018 International Conference on Engineering Technologies and their Applications (ICETA), Islamic University – ALNajaf - IRAQ

TABLE. I. COMPARISON AMONG THE ALGORITHMS BASED CMF

DETECTION TECHNIQUES.

Number of Publictions Per

50 38 42 33
Robustness 26 20
12 13 8 2 4

year
JPEG Compression
0

Computational

Gaussian noise
Illumination
Complexity
Author(s) Ref.

Rotation

Blurring
Scaling
, Year No. Algorithm Matching

Year
Babak Fig. 3 The No. of publications in the "IEEE" within the field of CMF detection
Invariant techniques between the years 2009-2018. Data were retrieved from
Mahdian Convent-
[25] Image High     "https://ieeexplore.ieee.org" with query "Copy-Move Forgery Detection".
et al., ional
Moments
2007
Invariant

Number of Publictions Per year

H.Hang et Ecludian
[17] key Less      54
al., 2008 Distance 60 45
points 38
40 28 22 24
X. Kang 18 16 13 15
et al., Convect-
20
[40] SVD Less      0
2008 ional

Invariant
S.Ryu et Convent-
[27] Image    
al., 2010 ional Medium
Moments
Year
E.
Texture
Ardizzone Edge Fig. 4 The No. of publications in the "Science Direct" within the field of CMF
[6] and less    
et al., Histogram detection Techniques between the years 2009-2018. Data were retrieved from
intensity "http://www.sciencedirect.com" with query "Copy-Move Forgery Detection".
2010
Invariant Best -Bin- V. CONCLUSIONS
X.Pan et
[36] key First  
al., 2010
points Algorithm This comprehensive display provided an overview of the
H.Solema Mutual Triangle recently existing CMF detection techniques for helping the
ni et al., [38] Informati distributio Less   researchers to develop their detection techniques with new ideas;
2011 on n via discussing the used algorithms, the challenges, pros, and
Yanping cons of the existing techniques. Where the core used, algorithms
Hang et
[21] DCT
Convent-
Medium   
are presented into different categories; namely, algorithms
al., 2011 ional based on DCT, algorithms using invariant image moments,
algorithms using texture and intensity descriptors, algorithms
Xiaomei Texture using invariant key points, algorithms based on mutual
Quan et [31] and K-NN   
less information, and algorithms based on SVD.
al., 2012 intensity Algorithm
Regrettably, in addition to the problem of detecting the
H.C.Hsu Texture
Gabor forgery localization, there is no technique can equip a perfectly
et al., [33] and Less  
2012 intensity
feature inclusive solution. Some techniques did not have a particular
robustness to common image processing processes such as
G.Zhang Invariant blurring, scaling, rotation, and jpeg compression. Also, some
Convect-
et al., [37] key High     techniques can be accurate but have high computational
ional
2012 points complexity.
S. REFERENCES
Mutual
Chakrabor Holistic
[39] Informati Less  [1] J. Waleed, Taha Mohammed Hasan and Thekra Abbas, "Comprehensive
ty et al., way Expansion in Anti-Forensics Techniques (AFTs) Based Compressed
on
2013 Image", Annual Conference on New Trends in Information &
Communications Technology Applications-(NTICT'2017), pp. 156-161,
Nathalie
Convent- March 2017.
Diane et [22] DCT High     [2] M. Fontani, Tiziano Bianchi, Alessia De Rosa "A Forensic Tool for
ional
al.,2013 Investigating Image Forgeries", International Journal of Digital Crime
and Forensics, Vol. 5, No. 4, pp. 15-13, October 2013.

159
2018 International Conference on Engineering Technologies and their Applications (ICETA), Islamic University – ALNajaf - IRAQ

[3] H. Ling, H. Cheng, Q. Ma, F. Zou, and W. Yan, "Efficient image copy in Electronics & Information Technology (PEIT), Alexandria, pp. 193-
detection using multiscale fingerprints," IEEE Magazine of Multimedia, 203, 2017.
Vol. 19, No. 1, pp. 60-69, 2012. [24] G. Liu, Junwen Wang, Shiguo Lian, Zhiquan Wang, "A passive image
[4] J. Waleed, Huang D. J. and Saad H., "An optimized digital image authentication scheme for detecting region-duplication forgery with
watermarking technique based on cuckoo search (CS)", ICIC Express rotation", Journal of Network and Computer Applications, Vol. 34, No.
Letters, Part B: Applications, Vol. 6, No. 10, pp. 2629-2634, 2015. 5, pp. 1557-1565, September 2011.
[5] J. Waleed, Huang Dong Jun, Sarah Saadoon, Saad H., Hiyam H., “An [25] B. Mahdian and Stanislav Saic, "Detection of copy-move forgery using
Immune Secret QR-Code Sharing based on a Twofold Zero a method based on blur moment invariants," Forensic science
Watermarking Scheme,” International Journal of Multimedia and international, vol. 171, no. 2, pp. 180-189, September 2007.
Ubiquitous Engineering Vol.10, No.4, pp.399-412, 2015. [26] J.-w. Wang, G.-J. Liu, Z. Zhang, Y. Dai, and Z. Wang, "Fast and robust
[6] J. Waleed, Huang Dong Jun, and Saad Hameed, “A robust Optimal Zero- forensics for image region-duplication forgery," Acta Automatica Sinica,
Watermarking Technique for Secret Watermark Sharing," International vol. 35, no. 12, pp. 1488-1495, Mar. 2009.
Journal of Security and Its Applications Vol.8, No.5, pp.349-360, 2014. [27] S.-J. Ryu, M.-J. Lee, and H.-K. Lee, "Detection of copy-rotate-move
[7] A. Dada Warbhe, Rajiv V. Dharaskar, Vilas M. Thakare, "Digital image forgery using zernike moments," in Information Hiding. Berlin,
forensics: An affine transform robust copy-paste tampering Heidelberg: Springer, pp. 51-65, 2010.
detection," 2016 10th International Conference on Intelligent Systems [28] N. Bakiah A., Ainuddin Wahid A. W., Mohd Yamani I. I., Roziana Ramli,
and Control (ISCO), Coimbatore, pp. 1-5, 2016. Rosli Salleh, Shahaboddin Shamshirband, Kim-Kwang R. C., "Copy-
[8] G. K. S. Gaharwar, Prof. V. V. Nath, R. D. Gaharwar, "Comprehensive Move Forgery Detection: Survey, Challenges and Future Directions",
Study of Different Types Image Forgeries", International Conference on Journal of Network and Computer Applications, Vol. 75, pp. 259-
Recent Advances in Engineering Science and Management, August 2015. 278, November 2016.
[9] Nikhilkumar P. Joglekar1 “A Compressive Survey on Active and Passive [29] H.Jen Lin, Chun-Wei Wang and Yang-Ta Kao, " Fast copy-move forgery
Methods for Image Forgery Detection “IJECS Volume 4 Issue 1 January, detection", WSEAS Transactions on Signal Processing, vol. 5, no. 5, pp.
pp.10187-10190, 2015. 188-197, 2009.
[10] AE Mohamed, MM Dessouky, KA Heba, FE Abd El-Samie, OA [30] Vivek Kumar Singh and R.C. Tripathi, "Fast and efficient region
Elshakankiry, OS Faragallah, “Comparative Study of Copy-Move duplication detection in digital images using sub-blocking method",
Forgery Detection Techniques “, International Journal of Innovative International Journal of Advanced Science and Technology, Vol. 35, pp.
Research in Computer and Communication, Vol. 5, No. 7, July 2017. 93-102, October 2011.
[11] J. A. Redi, W. Taktak and J. L. Dugelay, “Digital image forensics: a [31] Xiaomei Quan and Hongbin Zhang, "Copy-move forgery detection in
booklet for beginners”, Multimedia Tools Appl., vol. 51, no. 1, pp. 133– digital images based on local dimension estimation," Proceedings Title:
162, 2011. 2012 International Conference on Cyber Security, Cyber Warfare and
[12] J. Fridrich, D. Soukalm, J. Luka ´s ˇ, “Detection of copy-move forgery Digital Forensic (CyberSec), Kuala Lumpur, pp. 180-185, 2012.
in digital images,” Digital Forensic Research Workshop, Cleveland, OH, [32] J. Wang, G. Liu, H. Li, Y. Dai and Z. Wang, "Detection of Image Region
pp. 19–23, 2003. Duplication Forgery Using Model with Circle Block," 2009 International
[13] A. C. Popescu and H. Farid, “Exposing digital forgeries by detecting Conference on Multimedia Information Networking and Security, Hubei,
duplicated image regions,” Dept. Computer. Science, Dartmouth College, pp. 25-29, 2009.
Tech. Rep. TR2004-515, 2004. [33] H. C. Hsu and M. S. Wang, "Detection of copy-move forgery image
[14] W. Luo, Jiwu Huang and Guoping Qiu, "Robust Detection of Region- using Gabor descriptor," Anti-counterfeiting, Security, and
Duplication Forgery in Digital Image," 18th International Conference on Identification, Taipei, pp. 1-4, 2012.
Pattern Recognition, Hong Kong, pp. 746-749, 2006. [34] Devanshi Chauhan, Dipali Kasat, Sanjeev Jain, Vilas Thakare, “Survey
[15] G. Li, Q. Wu, D. Tu and S. Sun, "A Sorted Neighborhood Approach for On Keypoint Based Copy-move Forgery Detection Methods on Image”,
Detecting Duplicated Regions in Image Forgeries Based on DWT and Procedia Computer Science, Vol. 85, pp. 206-212, 2016.
SVD," 2007 IEEE International Conference on Multimedia and Expo, [35] A. Dada Warbhe, R. V. Dharaskar, V. M. Thakare, "A Survey on
Beijing, pp. 1750-1753, 2007. Keypoint Based Copy-paste Forgery Detection Techniques", Procedia
[16] A. N. Myna, M. G. Venkateshmurthy and C. G. Patil, "Detection of Computer Science, Vol. 78, pp. 61-67, 2016.
Region Duplication Forgery in Digital Images Using Wavelets and Log- [36] Pan, X., Lyu, S., "Detecting image region duplication using SIFT
Polar Mapping," International Conference on Computational features", 2010, IEEE Int. Conf. on Acoustics Speech and Signal
Intelligence and Multimedia Applications (ICCIMA 2007), Sivakasi, Processing (ICASSP), pp. 1706–1709, 2010.
Tamil Nadu, Vol. 3, pp. 371-377, 2007. [37] G. Zhang, H. Wang, “SURF based Detection of Copy-Move Forgery in
[17] H. Huang, W. Guo, and Y. Zhang, “Detection of Copy-Move Forgery in Flat Region,” International Journal of Advancements in Computing
Digital Images Using SIFT Algorithm,” in Proceedings of IEEE Pacific- Technology (IJACT) Vol. 4, No. 17, September 2012.
Asia Workshop on Computational Intelligence and Industrial [38] H. Soleimani, M. Khosravifard, "Mutual Information-Based Image
Application, Vol. 2, pp. 272-276, 2008. Template Matching with Small Template Size," 2011 7th Iranian
[18] S. Bravo-Solorio, A. K. Nandi, “Automated detection and localization of Conference on Machine Vision and Image Processing, Tehran, pp. 1-5,
duplicated regions affected by reflection, rotation and scaling in image 2011.
forensics”, Signal Proc., vol. 91, no. 8, pp. 1759–1770, 2011. [39] S. Chakraborty, "Copy move image forgery detection using mutual
[19] S. Kumar, P. Das, and S. Mukherjee, “Copy-Move Forgery Detection in information," 2013 Fourth International Conference on Computing,
Digital Images: Progress and Challenges,” International Journal on Communications and Networking Technologies (ICCCNT),
computer Science and Engineering, vol. 3, no. 2, pp. 652-663, 2011. Tiruchengode, pp. 1-4, 2013.
[20] O. M., Al-Qershi and B. E. Khoo, “Passive detection of copy-move [40] X. Kang and S. Wei, "Identifying Tampered Regions Using Singular
forgery in digital images: State-of-the-art,” Forensic Science Value Decomposition in Digital Image Forensics," 2008 International
International, vol. 231, no. 1, pp. 284–295, 2013. Conference on Computer Science and Software Engineering, Wuhan,
[21] Yanping Huang, Wei Lu, Wei Sun and Dongyang Long, “Improved Hubei, pp. 926-930, 2008.
DCT-based Detection of Copy-Move Forgery in Images”, Forensic [41] S. Bhosale, G. Thube, P. Jangam and R. Borse, "Employing SVD and
Science International, vol. 206, pp. 178-184, 2011. Wavelets for Digital Image Forensics and Tampering Detection," 2012
[22] N. Diane Wandji, Sun Xingming, and Moise Fah Kue, " Detection of International Conference on Advances in Mobile Network,
copy-move forgery in digital images based on DCT", IJCSI International Communication and Its Applications, Bangalore, pp. 135-138, 2012.
Journal of Computer Science Issues, Vol. 10, No 1, March 2013.
[23] M. A. Elaskily, H. K. Aslan, O. A. Elshakankiry, O. S. Faragallah, F. E.
A. El-Samie and M. M. Dessouky, "Comparative study of copy-move
forgery detection techniques," 2017 Intl Conf on Advanced Control
Circuits Systems (ACCS) Systems & 2017 Intl Conf on New Paradigms

160