Literature Survey On Digital Forensic by Using Image Mining: CH. Kodanda Ramu

You might also like

Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 11

Literature Survey on Digital Forensic by Using

Image Mining
CH. Kodanda Ramu
Assistant Professor, Department of CSE, Avanthi Institute of Engineering and Technology,
Tagarapuvalasa, Vizianagaram, Andhra Pradesh, India.

Dr. T. Sita Mahalakshmi


Professor, Department of CSE, GITAM Deemed to be University,
Visakhapatnam, Andhra Pradesh, India.

Abstract
Image mining is related to the development Keywords: Image handling, Image
of mining techniques in the field of digital Mining, Data Extraction, Information
forensics (DF). Image mining always works Preparing, Digital Forensic, Digital
on hidden data extraction and additional image, Acquisition, Machine learning.
information that is not clearly characterized 1. Introduction
in the pictures. The innovations in digital Digital Forensics (DF) is the emerging field
technology have new challenges now a days for identifying, retrieving and analyzing the
in the field of digital image forensics. In this part of evidence from digital data. Process
field, some traditional approaches and Diagram for Data Extraction is shown in
frameworks are proposed for proving the Fig1.1.
integrity and authenticity of digital images. Training Module Testing Module
Retrieval based tools are proposed for
recognizing unauthorized morphed copies of Image Acquisition Image Acquisition
the image. Authorization of a digital image
Pre-Processing Pre-Processing
can be proved by analyzing its
acquisition. In image mining, mining Feature Extraction Feature Extraction
process depends on several techniques like
Feature Clustering Feature Clustering
image handling, information preparing,
robotics and machine learning. The first Data Storage
imperative work of the mining is to get
immeasurably vital elements from the
concealed images. Image mining contains
diverse kind of uses in various segments like
research on space, remote detecting,
Comparison
medicinal finding etc. Image mining could
be a class of investigative systems that look
Image-Retrieval
at an outsized measure of image learning.
This paper introduces a review of various Output Image
image mining procedures.
Fig:1.1: Process Diagram for
Data Extraction
This small part of the evidence is one of the edited images. Almost all of these queries
crucial requirements in the cyber world. focus on the finding of the root or origin of
Image mining is part of the interdisciplinary the digital image with its capturing process
field of knowledge discovery in database either by camera or other sources. Evidence
management. Digital forensic is the wing of identified from such forensic research would
forensic science merged with recovery and give required forensic information for the
investigation of material found in digital use of cyber security, cloud security,
device due to an incident of computer crime intelligence agencies and law enforcement.
happenings. Digital forensics can be Details about image-acquisition techniques
classified into many which are shown in Fig: are useful in answering more forensic
1.2. questions to get confirmation about the
additional processing of the image. Image
mining is the strategy for seeking and
finding beneficial information and data in
enormous volumes of data. Image Retrieval,
Artificial Intelligence, Image preparing,
information mining are the techniques used
to assemble learning. These strategies
empower Image Mining to have
extraordinary methodology to separate data
from databases or accumulation of images
and in this manner, the option is to mine a
group of related character set information. In
design architecture of image mining and in
image processing, retrieving could be an
Fig:1.2: Classification Diagram of Digital uncommon sort of spatial property
Forensics for minimization. Information extraction
includes improving the quality of assets
Speedy developments in the field of digital
expected to clarify an outsized arrangement
photography are observed in recent years.
of data precisely. A few alternatives are used
Digital images are used in wide range of
in the Image retrieval framework. The
applications like medical diagnosis,
prevalent among them are Features in view
consumer photography, industrial
of shading, Features in light of surface and
photography, military, robotics etc. Due to
Features in light of shape. Image mining
high requirement with such a great demand,
strategy manages the extraction of certain
many low-cost images editing tools are
information and image with data related or
available in the market Due to that
diverse examples not explicitly available in
availability, many frauds and cybercrimes
the images. The image mining is to a great
are increasing in parallel with this tool
degree, particular because of the image
development, many forensic quires getting
database region unit dominantly non-
always crime or unauthorized data from
relative. Also, a few image credits don't
appear to be specifically obvious to the fleeting structures contained in SITS. Amid
client. Previously, image mining has been this approach, appalling or geometrical data
done to unravel differed issues together with was not thought about. This is the significant
target acknowledgment, visual observation, downside here.
confront acknowledgment, and face location Selim Aksoy and R. Gokberk Cinbi [3]
check. The objective of this paper is to audit proposed new directional spatial imperatives
image mining methods and procedures that which is the new image mining procedure.
are regularly valued. Image mining is related The creator utilized accuracy and physically
expansion of data mining. recognizes the items which fulfill each
Accordingly, image mining is a field of inquiry to figure the recovery execution. It
image processing and pattern matching with wants manual elucidation to process time.
data mining extensions. P. Rajendran and M. Madheswaran [4]
2. Literature Review specified a photo mining procedure. It joins
The writing overview in the investigation low level alternatives extricated from
of different image mining systems images and abnormal state data from
characterized some related innovation in the authority. It doesn't self-address the excess,
field of image handling. noisy image and time quality.
Sanjay T et al. [1] puts forward a photo Ankita Tripathi et al. [5] proposed a system
mining method utilizing wavelet change that classifies textual images (images that
display. distinguishing proof and encounter text within) using low-level image
information mining models are utilized as a features. Image classification and content-
part of the wavelet display. Discrete Wavelet based image retrieval is a growing field in
Transformation (DWT) + Principal the area of image classification. The
Component Analysis (PCA) framework is approach is based on various low-level
utilized as a model for framework image features including Gray Level Co-
recognizable proof. Taking care of sizable occurrence Matrix1 (GLCM) features like
measure of images at the supply machine mean, skewness, energy, contrast,
could be a critical assignment and it offers to homogeneity. Using these various features,
arise memory administration issue. The data the differences between images are
should be along these lines prudent to deal measured, and then these are used to classify
with images speedily. the textual images by performing
Lionel Lionel Gueguen and Mihai Datcu classification and clustering techniques on
[2] exhibited the data bottleneck rule to self- datasets. The proposed method
address the matter of separating applicable experimented on 60 different textual images
data from Satellite Image Test System to obtain an improved result that was not
(SITS). The model depends upon the obtained in earlier systems along with
properties analysis examination for the classification of images in three main
assurance of groups. Gauss– Markov categories: document, scene and caption.
irregular fields and auto binomial arbitrary Ross Brown et al. [6] outlined an image
fields is utilized to describe the spatial- mining framework for advanced measurable
applications. Bayesian systems are utilized may lead to misclassification sometime.
to change the data vulnerabilities that Hence an automated system is needed to
happen in scientific work. These are utilized solve our problem. The following major
to communicate between convictions, adjust categories of brain tumor images are taken
to entirely unexpected clients, diverse into our consideration. They are Metastatic
example recovers and copy human judgment bronchogenic carcinoma, Astrocytoma,
of semantic substance of image patches. Meningioma, sarcoma. The performance
Sanjay Silkari and Dr.Mahesh factor was evaluated against BRATS (Brain
Motwani [7] arranged a system which Tumor Segmentation) dataset. For the
concentrates on Color Moment Block purpose of calculating and extracting
Truncation Coding (BTC) which utilized various intensity related features MATLAB
shade is an element to extract the element tool is used. The experimental results
from image dataset. At that point, K-Means suggest that among the intensity-based
clustering equation was used to group the feature extraction methods GLCM (Gray
image dataset into various groups. Level Co-occurrence) method is showing
Tao Jiang and Tan [8] proposed two better results than the other methods.
approaches for finding the basic relationship Waikato Environment for Knowledge
among the content and images. The First Analysis (WEKA) tool classification
system finds the execution measures algorithm J48 also shows close correlation
between the information closeness and with GLCM Features.
visual choices. Another technique utilizes a Ja-Hwung Su et al. [11] presented a new
neural system to discover coordinate approach named Navigation-Pattern-based
mapping among visual and content Relevance Feedback (NPRF) by integrating
highlights by incrementally related the navigation pattern mining and a
highlights into a gathering of information navigation-pattern-based search approach
format. It must perform group learning on named NPSearch. The main feature of
the settled arrangement of unrefined NPRF is to efficiently optimize the retrieval
information. quality of interactive content-based image
Hemlatha and Devasana [9] proposed to retrieval (CBIR). As a result, traditional
search out the right image while mining a problems such as visual diversity and
media framework and built up a path for exploration convergence are solved.
mining images by recommending Lorenz Noorhaniza Wahid [12] arranged a swarm-
Information Measure (LIM) subordinate based recipe for order. He contrasted swarm-
image coordinating technique with the based procedure and Support vector
neural system. machine (SVM) and acquired the outcome.
B. Sathees Kumar and R. Anbu Selvi [10] Particle Swarm Optimization (PSO) must
focused on the comparison of three different distribute extra memory than Simplified
intensity-based feature extraction method for Swarm Optimization (SSO) for every
the abnormal patterns in brain tumors. molecule to acknowledge higher execution.
Physician's interpretation of brain tumors Satej Wagle et al. [13] proposed a medical
image classification model for retinal fundus
images. It uses a modified k-nearest image forensics takes, and, at the same time,
neighbor (KNN) algorithm, which uses an outperforms the former existing scheme.
instance-weighting scheme based on the Rajshree S. Dubey [16] delineated the
distance measure. The proposed model was image mining ways that rely upon the
implemented on “Normal” and “Severe” histogram and surface of the image. The
category of images. The model is evaluated required image is considered, the shading
bar diagram and Texture is made, the
using the performance metrics such as
resultant image was found. In this technique
classification accuracy and area under the
registering time was not considered.
ROC curve. The experimental results using Akshara Ravi et al. [17] studied about
the statistical texture-based features Recovering deleted files. When a file is
extracted from the images have shown that deleted, only pointers that link file's
the performance of this improved KNN metadata to its content are deleted and
model is significantly better than some of metadata entry is marked as deleted. As long
the other existing classifiers. as data is not overwritten or wiped, deleted
Surya S. Durbha et al. [14] created data will remain in unallocated space. One
highlight determination and highlight of the methods that can be used to recover
change in view of wrapper-based approach. these deleted files is file carving. File
It utilized locale-based framework. They carving reconstructs files only based on their
content unlike traditional data recovery
have contended that picking a significant list
methods that use metadata that points to the
of capabilities will expand the rate of
content. It is mainly done using headers and
semantic class. This framework directly
footers of file types. One of the primary
utilized symbolism from only one finder. challenges in file carving is to recover
Bambang Harjito and Heri Prasetyo deleted files that were fragmented. When a
[15] presents a new method on image file is fragmented, carving only based on
forensics application using the Error header and footer will produce corrupted
Diffusion Block Truncation Coding file. So, they discussed a method for carving
(EDBTC) feature. The image forensics tries fragmented document and image files from a
to detect the copy-move forgery image Universal Serial Bus (USB) drive.
regions on the forged image. Firstly, an . David A. Fay et al. [18] have built up a
image is divided into several non- model for multisensory image combination
overlapping image blocks. The image and intelligent image mining in light of
feature is further derived for each image neural models of shading vision picking up,
block. Herein, two image features, namely controlling and design acknowledgment.
Color Feature (CF) and Bit Feature (BF), are The modules like Image securing, Image
composed from the EDBTC compressed Fusion, extraction of Context highlights, and
data stream. The forged region is detected intuitive Image Mining produce a client to
while the image feature of this region is frame vector item in light of highlights like
similar to the image feature of another streets, streams, and woods and featured
region separated far away. As documented in target discoveries from crude multisensory
the experimental section, the proposed or multispectral symbolism. These yield
method gives a promising result on the images changed by recreated ecological
conditions not self-tended to in the memory, encoding and transformation of coefficients
for promoting identification. are the main steps in this process. Color
Herbert Daschiel and Mihai Datcu space conversion from Red, Green and Blue
[19] portray the thoughts of learning (RGB) to Y′ is the luma component and CB
substance-based image mining framework and CR are the blue-difference and red-
made to oversee and investigate monstrous difference chroma components (YCBCR) is
volumes of remote detecting image used for the transformation. The image is
information. The framework works both on divided into non-overlapping blocks to
the web and disconnected web interface. The decrease the number of operations. Then the
disconnected part goes for the extraction of image is transformed using integer discrete
crude image choices, their pressure, and data cosine transform (IntDCT). The transformed
decrease, the age of a totally unsupervised coefficients are then quantized using a
image content-record, and along these lines quantization matrix. After transformation,
the substantial procedure of the index the coefficients are 49 encoded. The
passage inside the data framework. This coefficients of a block are partitioned into
approach doesn't self-tend to the learning DC and AC coefficients and the DC
and data that is put away in the data coefficients are differentially encoded.
framework. Wynne Hsu and mong Li Lee Deepak Kumar Jain et al. [23] proposed a
[20] inspected the investigation issues in technique for image compression using
image mining, advancement in image discrete cosine transform (DCT) and
mining. They anticipated a data-driven adaptive Huffman coding (AHC). The
system for image mining. In that, they made original image is divided into blocks. DCT
out four levels of data: component level, is applied to each block by the pre-
Object level, semantics thought level, and multiplying the modified block with DCT
information level. This strategy does not matrix and by the post multiplying by
focus on further extension. transposes of DCT matrix. Each block is
Aura Conci and Everest then compressed through quantization. A
MathiasM.M.Castro [21] proposed a quantization matrix is used in combination
framework for mining images by color with a DCT coefficient matrix to carry out
content. The implementation was built using the transformation. The compressed image is
CGI, Java and C languages. The framework reconstructed through reverse process i.e.,
provides the possibility of use 5 distance by using inverse DCT. This technique has
function for evaluation of similarity among good performance as compared to other
images and 2 types of quantization. For algorithms.
comparing the influence on system Rajkumar and Latte [24] analyzed
performance of these parameters a renowned wavelet-based image encoding
simulating engine was implemented and all scheme (SPIHT). The Region of Interest
possible combinations have been tested (ROI) coding commences with the selection
considering sets of similar images. of ROI and its respective resolution by the
This framework can try online at user. The diverse ROIs are encoded with
html://www.caa.uff.br/~mathias/visual.htm. diverse resolution by applying lifting
wavelet transform (LWT) and SPIHT. Their
Tanima Dutta [22] proposed a new experimental results illustrate that using
technique for image compression. Image lifting wavelet transform and SPIHT, the
proposed ROI encoding scheme provides which they originated. This is timely work
high compression ratio and quality ROI. as new smartphone models large imaging
sensors, afford significant improvements in
Chenglong Chen et al., [25] Proposed classification rates using pattern noise. Our
Geometric transformations, such as resizing approach is to extract wavelet-based features
and rotation, are almost always needed when which are then classified using a support
two or more images are spliced together to vector machine. We show that this method
create convincing image forgeries. generalizes well when the number of source
cameras is increased.
Anil Dada Warbhe et al. [26] presented a Chao-Yung Hsu et al. [30] proposed an
digital image forensic method which can efficient cross-camera vehicle tracking
detect one of such image tampering. As technique via affine invariant object
images can be tampered in a number of matching. To improve the accuracy of Affine
ways, they address a common case called as and scale-invariant feature transform
copy-paste tampering. Their proposed (ASIFT) feature matching between images
method was robust to affine transform; from different viewpoints, They observed
especially to rotation and scaling. the spatially invariant property of ASIFT
and exploited this property with the min-
Sepideh Azarian-Pour et al. [27] proposed a hash technique to achieve efficient vehicle
new automatic method for discriminating object matching.
original and tampered images based on H. N. Gangavane et al. [31] presented a
“JPEG ghost detection” method, which is a model using new methodology for
subset of format-based image forensics evaluation of document clustering of
approaches. The inconsistency of quality criminal database by using k-means
factors indicates that the photo is a clustering technique. This model clusters the
composite one created from at least two criminal data basing on the type crime.
different cameras and therefore it is a Mandar Jadhav and K.K. Joshi [32]
manipulated photo. classification algorithm presented a methodology for acquisition of
first extracts the ghost border. Then the internal data from Firefox operating system
image is classified as original or tampered based mobile device and its complete
groups by thresholding a distance in feature analysis to recover important evidence. It
space. gives a way to start the process of acquiring
Neetu Singh and Rishabh Bansal [28] data by use of Android Debug Bridge,
studied Benford law in Digital Image followed by identification of important files
Forensics. Joint Photographic Experts Group which can be analyzed to obtain valid
(JPEG) and JPEG2000 approximately results. Further, it exhibits how use of
follow this law and JPEG2000 found closer Magnet axiom toolkit helped us to recover
to the given law. The non-following of this images, documents, audio files and web
law in different forensics setups can be used related data from the acquired evidence.
as fingerprint. Further, amount of deviation They also showed how tools like qtADB,
from Benford law in compressed images can FTK Imager and SQLite Forensics helped us
be used to find forgery. in forensic investigation.
Jocelin Rosales Corripio et al. [29] In this Mohammed Babiker et al. [33] Web
work we use the pattern noise of an imaging application attacks are an increasingly
sensor to classify digital photographs important area in information security and
according to the source smartphone from digital forensics. It has been observed that
attackers are developing the capability to might be appallingly useful. Encourage
bypass security controls and launch a large discoveries might be actualized later in the
number of sophisticated attacks. Several field of image mining. It is observed that
attempts have been made to address these many studies have been done on biometric,
attacks using a wide range of technology face, sketch art, iris, palm, etc. but
and one of the greatest challenges is researcher was not met high-efficiency range
responding to new and unknown attacks in due to improper information on input. So,
an effective way. This study aims to this work can be carried out further by using
investigate the techniques and solutions used efficient digital forensics techniques.
to detect attacks, such as firewalls, intrusion
detection systems, honeypots and forensic
techniques. Data mining and machine
learning techniques, which attempt to 4. References
address traditional technology shortcomings [1]
and produce more effective solutions, are Sanjay T. Gandhi, K. T. Talele and Avinash
also investigated. It was aimed to contribute G. Keskar, “Image Mining Using Wavelet
to this growing area of research by exploring Transform”, Knowledge-Based Intelligent
more intelligent and convenient techniques Information and Engineering Systems,
for web application attack detection by Springer link book chapter, pp. 797-803,
focusing on the data mining techniques in 2007.
forensics.

3. Conclusion
This paper exhibits a review of various [2]
image mining procedures that was proposed Lionel Gueguen and Mihai Datcu, “Image
in past years. It conjointly conveys Time-Series Data Mining Based on the
calculated synopsis of the system. Image Information-Bottleneck Principle”, IEEE
mining is only the learning mining inside the Transactions on Geoscience and Remote
field of image processing and pattern Sensing, Vol. 45, no. 4, 827, April 2007.
recognition. The future examination work
could grasp the usage of the Bayesian [3]
system for connectedness inputs and a lot of Selim Aksoy and R. Gokberk Cinbi, “Image
serious tests with elective specimens of Mining Using Directional Spatial
image unauthorized work. It is also Constraints,” IEEE Geosciences and Remote
visualized that subjective testing is Sensing Letters, vol. 7, no. 1, pp. 33-37, Aug
performed with contribution from scientific
2009.
authorities. Some feasible fates think about
which will be directed inside the space of
[4]
image mining grasp the experimentations on
P. Rajendran and M. Madheswaran, “An
elective image parts like surface, shape, so
Improved Image Mining Technique for
forward. It additionally will interest to
Brain Tumor Classification Using Efficient
investigate hidden connections among
Classifier”, International Journal of
images. For example, Intensive and serious
Computer Science and Information Security,
beta example investigation worried inside
Vol. 6, no. 3, pp. 107-116, Jan 2009.
the current frameworks in the database
[5] Ja-Hwung Su, Wei-Jun Huang, Philip S. Yu,
Ankita Tripathi, Shivam Pandey, Hitesh and Vincent S. Tseng, “Efficient Relevance
Jangir,” An improved and efficient image Feedback for Content-Based Image
mining technique for classification Retrieval by Mining User Navigation
of textual images using low-level Patterns”, IEEE Transactions on Knowledge
image features“, Inventive Computation and Data Engineering, vol. 23, no. 3, March
Technologies (ICICT) 2017. 2011.

[6] [12]
Ross Brown, Binh Pham, and Olivier de Vel, Noorhaniza Wahid, “A Novel Image
“Design of a Digital Forensics Image Classification Algorithm Using Swarm-
Mining System”, Springer-Verlag Berlin Based Technique for Image Database”,
Heidelberg, pp. 395.404 Sept 2005. Springer-Verla-Berlin Heidelberg, pp.
460–470, April 2011.
[7]
Sanjay Silakari, Mahesh Motwani, and [13]
Manish Maheshwari “Color Image Satej Wagle, J. Alamelu Mangai and V.
Clustering using Block Truncation Santosh Kumar,” An Improved Medical
Algorithm”, Ijcsi International Journal of Image Classification Model using Data
Computer Science Issues, Vol. 4, no. 2, Aug Mining Techniques”, 2013 IEEE GCC
2009. Conference and exhibition, November 17-
[8] 20, Doha, Qatar.
Tao Jiang and Ah-Hwee Tan, Learning, [14]
“Image-Text Associations”, IEEE Surya S. Durbha, Roger L. King, and
Transactions on Knowledge and Data Nicolas Yunnan, “Wrapper-Based Feature
Engineering, Vol. 21, no. 2, February 2009 Subset Selection for Rapid Image
Information Mining”, IEEE Geoscience and
[9] Remote Sensing Letters, vol. 7, no. 1,
M. Hemalatha and C. Lakshmi Devasena, January 2010.
“A Hybrid Image Mining Technique using
LIM based Data Mining Algorithm”, [15]
International Journal of Computer Bambang Harjito and Heri Prasetyo ,”
Applications, Vol. 25, no.2, pp. 1-5, July Image forensics using EDBTC feature”,
2011. Consumer Electronics - Taiwan (ICCE-TW),
2017 IEEE International Conference.
[10]
B. Sathees Kumar, R. Anbu Selvi,” Feature [16]
extraction using image mining Rajshree S. Dubey, “Image Mining using
techniques to identify brain Content-Based Image Retrieval System”,
tumors“, Innovations in Information, (IJCSE) International Journal of Computer
Embedded and Communication Systems Science and Engineering Vol. 02, no. 07, pp.
(ICIIECS), 2015 International Conference. 2353-2356, October 2010.

[11] [17]
Akshara Ravi, T Raj Kumar, Angelo Renju Emerging Trends & Technology in
Mathew,” A method for carving Computer Science (IJETTCS) , Volume 3,
fragmented document and image files Issue 1, January – February 2014.
“Advances in Human Machine Interaction
(HMI), 2016 International Conference. [24]
T. M. P. Rajkumar and Mrityunjaya V
[18]
Latte.” ROI Based Encoding of Medical
David A. Fay, Richard T. Ivey, Neil
Images: An Effective Scheme Using Lifting
Bamberger, and Allen M. Waxman, “Image
Fusion & Mining Tools for a COTS Wavelets and SPIHT for Telemedicine”
Environment”, ISIF pp. 606-613, 2003. International Journal of Computer Theory
and Engineering, Vol. 3, No. 3, June 2011.
[19]
Herbert Daschiel and Mihai Datcu, [25]
“Information Mining in Remote Sensing Chenglong Chen, Jiangqun Ni,Zhaoyi
Image Archives: System Evaluation”, IEEE Shen,” Blind Forensics of Successive
Transactions on Geoscience and Remote Geometric Transformations in Digital
Sensing, Vol 43, no. 1, January 2005. Images Using Spectral Method: Theory and
Applications “ IEEE Transactions on Image
Processing ( Volume: 26, Issue: 6, June
[20] 2017 ).
Wynne Hsu, Mong Li Lee, and Ji Zhang, [26]
“Image Mining: Trends and Developments”, Anil Dada Warbhe, Rajiv V. Dharaskar,
Journal of Intelligent Information Systems, Vilas M. Thakare ,”Digital image
Vol. 19, no.1, pp.7-2, Jan 2002. forensics: An affine transform robust
copy-paste tampering
[21] detection“, Intelligent Systems and
Aura Conci and Everest Mathias M. M. Control (ISCO), 2016 .
Castro, “Image mining by content”, Expert
Systems with Applications, Vol. 23, no. 4, [27]
pp. 377-383, April 2002. Sepideh Azarian-Pour, Massoud Babaie-
Zadeh, Amir Reza Sadri,” An automatic
[22] JPEG ghost detection approach for digital
Tanima Dutta, “Medical Data Compression image forensics “ Electrical Engineering
and Transmission in Wireless Ad Hoc (ICEE), 2016.
Networks,” IEEE Sensors Journal, Vol. 15,
[28]
No. 2, Feb. 2015.
Neetu Singh and Rishab Bansal, “Analysis
of Benford's law in digital image
[23]
forensics“Signal Processing and
Deepak Kumar Jain, Devansh Gaur , Kavya
Communication (ICSC), 2015.
Gaur , Neha Jain “Image Compression using
Discrete Cosine Transform and Adaptive [29]
Huffman Coding” International Journal of
Jocelin Rosales Corripio, David Manuel
Arenas González, Ana Lucila Sandoval
Orozco,” Source smartphone
identification using sensor pattern noise
and wavelet transform.“ Imaging for
Crime Detection and Prevention, 2013.

[30]
Chao-Yung Hsu, Li-Wei Kang, Hong-Yuan
Mark Liao,” Cross-camera vehicle tracking
via affine invariant object matching for
video forensics applications
“ Multimedia and Expo (ICME), 2013 IEEE
International Conference.

[31]
H. N. Gangavane, M. C. Nikose, P. C.
Chavan,” A novel approach for document
clustering to criminal identification by
using ABK-means algorithm“ Computer,
Communication and Control (IC4), 2015
International Conference.

[32]
Mandar Jadhav, K.K. Joshi,” Forensic
investigation procedure for data
acquisition and analysis of Firefox OS
based mobile devices.“ Computing,
Analytics and Security Trends (CAST)
2017.

[33]
Mohammed,Babiker,EnisKaraarslan,Yasar
Hoscan,"Web Application Attack Detection
and Forensics: A Survey",978-1-5386-
3449-3/18/$31.00 ©2018 IEEE

You might also like