Theses supervised by Yrd. Doç. Dr. Reza Zare Hassanpour
10 theses · Çankaya University
Analyzing digital image manipulation detection algorithms
In recent years tampering in digital image is becoming very widespread and more interesting. With powerful Image software, editing tools make a tampered image easy, thus the field of image forensic has become very important and crucial. In this thesis, investigation is made on the image by depending on the traces that occur when the soft correlation between rows and columns are divested in altered images, in term of linear dependency and exploitation of these traces in image forensic, depending on the properties of singular value decomposition and derived features which can be helpful and accurate sufficiently to reveal the tampering that is done on the genuine digital image.
Manipulation detection in uncompressed video
In recent years due to advancement in video and image editing tools it has become increasingly easy to modify content of multimedia. Doctored videos are very difficult to identify through visual examination as artifacts left behind by processing steps are subtle and cannot be easily captured visually. Therefore, the integrity of digital videos cannot be taken for granted and these are not readily acceptable as a proof-of-evidence in a court-of-law. Hence, identifying the authenticity of videos has become an important field of information security. This thesis presents an approach to detect and temporally localize video forgery, based on the correlation of noise residual using the Discrete Wavelet Transformation (DWT) algorithm for de-noising. The proposed algorithm is tested on public datasets such as SULFA are used for performance evaluation. The results show that the approach is effective against manipulation techniques. In addition, it detects and localizes tampered frames in a video with high accuracy.
Liver volume computation from computed tomography (CT) images.
Recently, the computed tomography images have been used explicitly for diagnosing of liver disease, as well as, they are used to measure the volume of the treatment and relocate of liver. In general, the liver is segmented by the medical images using manual or the computed tomography segmenting techniques because the characteristics surface of the liver and the effects of the partial volume makes the automatic discernment of liver from the other nearby organs very complicated process. Consequently, in this thesis we will suggest a semi-automatic technique for the liver for the purpose of measuring the size of the liver by using the tomography computed images. Thus, in order to segment the area of interest, we have to focus on the process of segmenting the liver based on the edge in addition to the techniques of the growing area. Finally, the investigational measurements of the size and area of the liver are consistent with those use for manual tracking methods for radiographic doctor and it shows that this algorithm is an effective algorithm to measure the size and semi-automated segmentation of the liver.
Scalable, secure and interoperable design for the Internet of Things
A vast number of "things" have been used to meet innumerable commitments of ICT in our contemporary world. Most of these "things" are placed in different locations and regularly work to serve individual users or groups and give them the ability to access them by using their own networking, applications and/or individual databases. Most of these applications run these data exclusively. For these reasons, it will not be easy to integrate a third party application within them or ensure the security and privacy of collected data. Moreover, end users face difficulties in accessing these data from anywhere in a unified form. At the present moment, the adoption of delivering this technology in new ways is growing rapidly. In this thesis, a new design for the Internet of Things is proposed, which we call "Web IoT." The new design can play a role between things and stakeholders by virtualizing these entities and making them available to end users from anywhere. It also offers a wide range of tools and options for stakeholders to host many types of entities in one place, thereby giving them the ability to control and manage content, share this content in social networks, apply more personalization, dynamically update, and much more. Our design was tested in terms of scalability, functionality and flexibility. Web IoT will help to overcome the limitation of collected data in a secure and scalable manner.
Developing preventive mesurements against attacks on wireless networks
Wireless Sensor Networks (WSN) and Ad Hoc networks are being using in different applications and fields of our life in the modern era. They are used in the military, medical, industrial, disaster relief, and commercial applıcatıons. Security is an important challenge in the wireless networks. Most of the wireless networks are susceptible to different attacks. In this study we address Sybil attack which uses multiple identities for a node to create the illusion that these identities come from multiple nodes. In this thesis, we propose a method to detect and prevent Sybil attacks. In our method, we consider WSN application in vehicular networks (VANET) where we detect Sybil attack without knowing the identity of the malicious node which preserves the user privacy as well. The proposed method is simulated in NS2, a widely used open source network simulator. The results are compared with the realistic test case conducted by the centralized authority DMV in terms of the detection rate and overheads.
Quantitive management of information security in organizations
Different methodology implementation to define quantitive management of information security in organizations including defining information risks quantitive approach using Lenstra and Voss [1] suggestion, annual loss expectancies like model to meet the expectations of the real world applications like cost management, finance management etc. We made some improvements on approach using ISO/IEC 27005:2011 framework. It was chosen because of a global standard and included in ISO 31000:2009, also my real life experiences deal with ISO/IEC 27001 implementation and certification.
Developing an efficient stream control transmission protocol (SCTP) multi-streaming using pluggable priority algorithm for network optimization
This thesis introduces a pluggable priorities algorithm for the SCTP protocol(PP-SCTP) as a method of reducing delay in the transmission of main data during periods of low bandwidth access. The SCTP protocol in the transport layer uses a Round-Robin or First-Come, First-Served method to transfer the data. These methods do not provide efficient and flexible choices. We set the priority of the streams in accordance with client demand. The PP-SCTP is useful for applications that send different types of data, such as still images, video, text, and documents. The goal of this thesis is to obtain a more flexible and more efficient choice for the end user or customer. The problem germane to this thesis lies in the transport layer, examples of which include the SCTP protocol with multi-streaming. Coding was written to test the various combinations and permutations of algorithms and protocols. In Chapter 5, we discuss the results of our work with the PP-SCTP priority algorithm, most notably its greater efficiency and the reduction of average latency in comparison to other algorithms and protocols in various combinations and permutations, including the First-Come, First-Serve data transfer method.
Blind linear correlation technique for image watermarking
In digital environment, make, change, update, distribute and store digital data are convenient, therefore as much abuse of digital data is added. This calls for a method to prove the ownership right on digital contents and to avoid unauthorized users to tamper and distribute digital data. Thereby to achieve this security requirement, watermarking schemes are introduced that have applications in all three forms of media, i.e., video, music and image. This thesis aims to test blind linear correlation technique by Stirmark benchmark 4.0 that contains sixteen different tests, such as attacks and distortions which are well-known in image processing. Those tests are applied on the images which are watermarked by blind linear correlation technique. In this thesis, results of this experiment are discussed and analyzed.
Detecting location of hidden messages on digital images using RS steganalysis method
Steganography is a branch of methods that deal with hiding information in cover media. The result of the hiding process should resist detection by any means. To detect the hidden information, several analysis methods have been developed (named as steganalysis methods). Steganalysis of Regular and Singular Groups (RS) is a method which aims to detect hidden information on digital images. This method is helpful for estimating length of the hidden message but it does not find over which section of the image the hidden message is. The aim of this thesis is to locate in which part(s) of the image the message is hidden by utilizing this steganalysis method and quadtree data structure.
Analysis of the loads and difficulties in IPv4 to IPv6 transition
This thesis focuses on analysing of the loads and difficulties in IPv4 to IPv6. Due to the depletion of the IPv4 addresses, IPv6 transition is inevitable. Therefore all parties especially network providers, ISPs (Internet Service Providers), hardware manufacturers, software developers and goverments need to prepare itself to this transition. In this thesis obligations of all parties before and after the transition and difficulties of the transition are described in details.