Show: 25 50 75 100 Results

Search results: 25 out of 130

محاكاة التطبيق العسكري الامن استنادا على شبكات الراديو الادراكية == Simulation of Secure Military Application based on Cognitive Radio Networks

Author name: امير سمير حمود محمد علي
Supervisor name: ستار بدر سدخان
General topic: Computer Science
Specific topic: Network Security
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

كشف الكائن الفيديوي المخفي باستخدام تعقب متوسط الازاحة وتقنيات الخياطة == Video Hidden Object Recovery Using Mean Shift Tracking and Stitching Techniques

Author name: قاسم جليل خضير عبيس
Supervisor name: اسراء هادي علي
General topic: Computer Science
Specific topic: Software
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

تقنية جدولة فعاليات عقد الاستشعار الحافظة للطاقة في شبكات الاستشعار اللاسلكية

Author name: وسام حسين صالح عبد
Supervisor name: علي كاظم ادريس
General topic: Computer Science
Specific topic: Computer Networking
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

اضافة العلامة المائية للفيديو اعتمادا على مسار حركة الكائن == Video Watermarking Based on Object Motion Trajectory

Author name: صفا سعد عباس المرعب
Supervisor name: اسراء هادي علي
General topic: Computer Science
Specific topic: Computer Science
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

تحسين نظام التوصية التعاونية بالاعتماد على ومعلومات

Author name: زينب خير الله كاظم
Supervisor name: هدى ناجي نواف المعموري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

اخفاء صوره داخل صوره بالاعتماد على FMM == Image in Image Steganography Based on FMM

Author name: حسين علي حسين
Supervisor name: بهيجه خضير شكر
General topic: Computer Science
Specific topic: Image Processing
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

كشف التعديلات الخبيثة على الملفات التنفيذية المحمولة في بيئة الشبكة == Malicious Modification Detection of Portable Executable Files in Network Environment

Author name: مصطفى عبد الرسول علي
Supervisor name: وسام سمير بهيه
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

تميز الارقام المطبوعة باستخدام نظام اندرويد == PRINTED NUMBER RECOGNITION USING ANDROID SYSTEM

Author name: قاسم محمد جاسم
Supervisor name: وسام سمير بهيه
General topic: Computer Science
Specific topic: Information Technology
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

توليد قواعد المنطق المضبب بالاعتماد على خوارزمية الاختزال و طريقة الانحدار == Fuzzy Rule Generation based on Subtractive Clustering and Gradient Descent

Author name: زهراء عبد محمد
Supervisor name: حسين عطية
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

طريقة استخلاص و تمييز النص الاصطناعي في ملف الفيلم == Extraction and Recognition Method for Artificial Text in Movie File

Author name: مريم حسين محمد بحر
Supervisor name: توفيق عبد الخالق الاسدي | اسراء هادي علي الشمري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

مطابقة السلاسل المتعددة بالاعتماد على الخوارزمية الجينية المطورة == Multiple Sequence Alignment Based on Developed Genetic Algorithm

Author name: فنار عماد خزعل الخزاعي
Supervisor name: نبيل هاشم كاغد | ايمان صالح الشمري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

تعديل مسار كائن استنادا على مسار مثالي في الفيلم == Object’s Trajectory Modification Based on Typical Trajectory in Movie

Author name: سارة عبد الرضا عبد
Supervisor name: اسراء هادي علي الشمري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

تقنية تقسيم بيانات لتحسين خوارزمية العنقدة DBSCAN == Data Partitioning Technique to Enhance DBSCAN Clustering Algorithm

Author name: اسراء صالح كامل زريزب
Supervisor name: صفاء عبيس المعموري
General topic: Computer Science
Specific topic: Algorithms
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

تحسين تفسير قواعد الارتباط بالدمج بين التعميم وطريقة العرض المعتمدة على المخطط == Enhancement of Association Rules Interpretability by Combining Generalization and Graph - Based Visualization

Author name: زهراء نجم عبد الله مهدي
Supervisor name: صفاء عبيس المعموري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

العنقدة اعتمادا على العقدة المتطرفة لنقل البيانات في شبكاث الاستشعار اللاسلكية == Extreme - Node Clustering for Data Transmissions in Wireless Sensor Networks

Author name: حوراء عبد الكاظم حسن
Supervisor name: سعد طالب الجبوري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:

الاداء الامثل لشبكات الاستشعار اللاسلكية باستخدام نماذج نشر مختلفة == An Optimal Performance of the Wireless Sensor Networks Using Various Deployment Models

Author name: عبد الناصر رياض فنجان سالم
Supervisor name: سعد طالب حسون الجبوري
General topic: Computer Science
Specific topic: Software
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: Wireless sensor networks (WSNs) today are widely used in various military and civilian applications and in the construction of a new concept called Internet of thinks (IoT), so it has been given great importance especially in recent years. In most WSNs applications sensors are deploying in random manner. Such randomness deployment produces trivial control on the network with no coverage guarantee and may achieve weakly connected network topology. Therefore, precise location can be often pursued for different nominated applications with the aim of configuring the topology of the network to reach to the requirements of preferred application. Most of the important WSN optimization techniques are to place the sensors in a deterministic manner to meet the required performance aims.In this dissertation several solutions are proposed to handle the deployment problem in WSN such as coverage, connectivity and reliability. Our suggestion depends on developing certain re - deployment approaches. These approaches are suggested and implemented in two virtual phases. In the first phase a random deployment was suggested then improves the locations of all the deployed sensors in the second phase. Once completing these phases a set of feasible locations will be available and can be used in the process of real sensor deployment.Seven algorithms are suggested, two of them were developed from the centralized optimization algorithms. The first is Particle Swarm Optimization (PSO), one of the common optimization methods. And the second is called Grid - distribution, where we considered this model to estimate the optimal number of sensors needed to cover a specific area. On the other hand, we have extracted two new models from Grid - distribution fundamentals, one of them to cover the border called barrier deployment, and the second is to cover center a certain area called center deployment.Three other optimization deployment algorithms are proposed to redeploy sensors after initial random deployment and improve coverage, connectivity and network reliability at the minimum cost. These algorithms are Distance based deployment, Markov based deployment and Angles based deployment. A mathematical model has been built for each of these proposed algorithms and has been implemented and tested by the Net Logo simulator. Each algorithm is executed in all sensors to achieve the desired objectives. The results of these algorithms were shown to be superior and dominate the results of existing algorithms such as Glowworm Swarm Optimization (GSO) and PSO_Voronoi.

اقتراح خوارزمية هجينة للتشفير الكتلي == Proposed Hybrid Block Cipher Algorithm

Author name: احسان احمد محمد لهمود
Supervisor name: عبد الكريم عكلة عبادي
General topic: Computer Science
Specific topic: Computer Science
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: يعتبر التشفير من المجالات الجيدة في الوقت الحالي كما نعلم ان الامن شرط اساسي لاي عمل ومن اجل ذلك نحن بحاجة الى خوارزمية قوية جدا وغير قابلة للكسر لتوفير اجراءات امنية مشددة.لذلك نحن نحتاج الى خوارزمية للتشفير وفك التشفير لتوفير امنية عالية جدا وانتاجية جيدة جدا. اذا نظرنا الى العالم الحقيقي، هناك الكثير من المنظمات التي لديها قاعدة بيانات كبيرة جدا مع اجراءات امنية مشددة. وفقا للقلق الامني، تعمل بعض خوارزميات التشفير وفك التشفير لحماية المعلومات السرية مثل DES و3DES وAES وBlowfish.تم اقتراح وتصميم خوارزمية هجينة لتشفير كتلة او لفك تشفيرها مكونة من 256 بت باستخدام مفتاح بطول 288 بت. يتم تحويل كتله بطول 32 - حرف من النص الواضح او النص المشفر الى 256 بت. يتم جدولة المفتاح السري لكي يم تطبيقه في عملية التشفير وفك التشفير. يتم اخضاع كتلة النص الواضح الى عملية التقلب الاولية، وفي نهاية التشفير يتم اخضاع النص المشفر الى التقليب النهائي. تم تصميم الخوارزمية المقترحة للدمج بين اثنين من الخوارزميات (على اساس فيستيل وغير فيستيل).استخدمت في هذه الاطروحة بعض من معاير التشفير الكتلي مثل الانتاجية لتوليد كتلة مشفرة حيث حققت انتاجية الخوارزمية المقترحة قيمة 27.240 كيلوبت في الثانية. اما بالنسبة لهجمات القوة الغاشمة حيث تحتاج 1079 X 1.57سنة اذا تم تطبيقها لمهاجمة مفتاح الخوارزمية ، حققت الخوارزمية المقترحة نسبة اكثرمن ٥٠% ضمن معيار SAC حيث كانت النسبة (٥١.١٧%) وكذلك بالنسبة لمعيار BIC حيث حققت نسبة (٥٣.١٢%). تم تنفيذ الخوارزمية المقترحة باستخدام لغة البرمجة (Microsoft Visual Basic.Net 2008) وعلى حاسوب ذو مواصفات (Windows 10 pro, processor : Intel(R) core (TM) i7 - 3612QM CPU @ 2.10GHz, RAM 6.00 GB, and system type : 64 - bit operating system). | The Cryptography is very good area for research now a days. As we know that security is very primary requirement for the any business. And we need very strong and unbreakable algorithm which provides high security. We need encryption and decryption algorithm which is having very high security with very good throughput. If we look at the real world, lots of organizations are having very large database with high security. Some encryption and decryption algorithms are working behind confidential information like DES, 3DES, AES and Blowfish.A proposed hybrid algorithm designed to encrypt or decrypt block of a message that consisting of 256 - bit with control of a 288 - bit as a key length. The blocks constructed by converting a 32 - charecter block of plaintext or ciphertext into 256 - bit. The secret key is scheduled to be applied to encrypt and decrypt. Plaintext block will be subjected to an initial permutation IP, and final permutation. The proposed algorithm designed in a fashion which belongs on two algorithms (based on Feistel and Non - Feistel). In this dissertation, some components used like throughput of generate encryption block. It has achieved as 27.240 Kbps. Based on brute force attacks may be applied on this algorithm where it needs 1.57x1079 years to attack the applied key, the security is provided in this algorithm achieved results more than 50% within criteria of SAC is (51.17%) and BIC is (53.12%). The proposed algorithm were implemented using the programming language (Microsoft Visual Basic.Net 2008) within computer information of (Windows 10 pro, processor : Intel(R) core (TM) i7 - 3612QM CPU @ 2.10GHz, RAM 6.00 GB, and system type : 64 - bit operating system)

تحسين خوارزمية تشفير A5/1 بالاعتماد على تقنية الترشيح لتطبيق انترنيت الاشياء == Improvement of A5/1 encryption algorithm based on filtration technique for IoT Application

Author name: زينب حمزة جاسم
Supervisor name: ستار بدر سدخان المالكي
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: The Internet of Things (IoT) is an environment in which people, animals or objects are equipped with individual identities and have the ability to transmit data across the network without the need for human - to - human or human - to - computer interaction.Security of IoT is very important because it is used in many fields and will have a big impact on the IoT industry. Internet of Things is similar to conventional computer networks, so security requirements such as confidentiality, integrity, availability and non - repudiation must be taken into account in building a network environment. One of appropriate solutions for providing security in IoT is cryptography. At present, traditional cryptography solutions focus on producing high levels of security, but they are slow in speed, large in size and consume a loT of energy, ignoring the conditions of constrained devices that use in IoT. These devices require appropriate cryptographic algorithms to suitable their characteristics, and this considers as big challenge.In this thesis, we propose a modification of the A5 /1 stream cipher to Internet of Things (IoT) by adding a fourth register and applying a filtration function on registers to increase the linear complexity of the algorithm and to strengthen the linear combination function (XOR) with remain total number of registers is 64 - bit. A5/1 Stream cipher is considered an efficient implementation of hardware, but insecure for use in such IoT applications. So we take efficiency implementation of hardware of A5/1 to produce proposed algorithm with the highest security, also efficient in hardware, and suitable for application in resource devices environments such as the Internet of Things. To make sure that the proposed algorithm as lightweight and can be applied to the constrained devices used within IoT, we have

طريقة تحليلية لانظمة التشفير الصوتي البايومتري == Analytical Approach of Biometric Based Voice Encryption System

Author name: علي كاظم مطر
Supervisor name: ستار بدر سدخان المالكي | بهيجة خضر شكر
General topic: Computer Science
Specific topic: Computer Science
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: There is no absolute security for important systems can ever get because attackers have always the ability to broke and attack them through their disadvantages. These days, biometrics is used to raise protection rate compared with the traditional methods. Authentication systems and cryptosystems are some of the important aspects of the practical life that use biometrics. The biometric voice is one of these traits that can build suitable secure systems according to the varying in the biometric voice.In this dissertation, two proposed analyzers were introduced to analyze Authentication system of biometric voice and the encryption process of the cryptosystems. The Authentication analyzer was used Additive White Gaussian Noise (AWGN) to stimulate effect of background/ transmission channel noise, and analyze the behaviors of the authentication system using FAR and FRR biometric performance measures.Performing proposed Authentication analyzer found degradation in the accuracy about (9.6% to 19.2%) in Splashdata database, about (1.9% to 5.7%) in Texas Instrument of Massachusetts Institute and Technology (TIMIT) database, and 0% in Texas Instruments - Digits (TIDIGITS) database of all selected members cannot be rejected illegally even when AWGN was reached 20 dB. SNR. Also, performing proposed Authentication analyzer found that the highest security degradation was about 0.05769 in Splashdata database, 0.01923 in TIMIT database and 0.05769 in TIDIGITS database of all selected members even when AWGN was reached 20 dB. SNR.The second proposed analyzer (Encryption analyzer) was also tested on the same databases to analyze them according to three randomness tests from National Institute of Standards and Technology (NIST) packages (pre - encryption phase), then these databases were also tested using Cross - correlation and Chi tests (post - encryption phase). These two phases produced two results by implementing two Mamdani fuzzifiers to get the certainty of each result. The final output of proposed Encryption analyzer was produced by merging the two previous Mamdani fuzzifiers in another final Mamdani fuzzifier.The performance of the Encryption analyzer was proved by classifying good/bad Keystream using the (post - encryption tests) and also with the comparison of the average value of other different 5 randomness tests from NIST.Finally, the ANFIS structure was used to generalize the hidden relationships that trained from the three randomness tests (pre - encryption process). The generalization process made ANFIS having the ability to predict the values of unseen (untrained) patterns. ANFIS results were promising results according to the proposed Encryption analyzer.Chapter

تحسين نظام مقترحات هجين باستخدام خورازمية الترتيب PageRank == Hybrid Recommender System Enhancement Using Personalized PageRank Algorithm

Author name: حیدر مجید ناجي
Supervisor name: غیداء عبد الحسین بلال
General topic: Computer Science
Specific topic: Information Technology
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: Recommendation system is an information filtering system. PageRank Algorithm is useful method for the recommendation task. PageRank algorithm is used to improve the representation of movies and users in the graph network in addition to ranking movies for a target user in the recommendation activity. User preferences has been calculated based on movies genres as a part of building the user profile from his past history of rating (content analysis), and for each user, this preference was used to produce a personalized initial pagerank value for each movie instead of the traditional static equal initial pagerank value which is the first stage of iterative personalize pagerank algorithmIn addition to personalize the target user, new way has been presented to personalize each user with different weight according to them rating on movies, by supporting the proposed recommender system with a new personalize parameters for each user, which is the manner of the collaborative filter approach.To evaluate the performance of the proposed recommender system and measure the accuracy of recommendations, precision, recall and F - Measure metrics has been used. MovieLens2K dataset used to evaluate the recommendation system, it has 10,197 movies and 2,113 users, in average of 85 movie rating per user.The experimental results showing that, there is a significant improvement for the recommendations process, which is mean hybrid recommender system using personalized pagerank (HRS - PPR) is better than the traditional recommender system using personalized pagerank.

عنقدة الصور اعتمادا على طريقة كسورية مطورة وتنقيب المخططات == Image Clustering Based on Developed Fractal Method and Graph Mining

Author name: فراس صبار مفتن
General topic: Computer Science
Specific topic: Computer Science
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: تشير عنقدة الصور الى تقسيم الصور الى عدة مجاميع. حيث كل مجموعة تسمى عنقود حيث يحتوي على صور متشابه في الخصائص ولكنها مختلفة عن الصور في العناقيد الاخرى. يمكن تفسر الخصائص الشاملة كميزات احصائية على انها خاصية للصورة تشمل جميع وحدات البكسل المستخدمة لحساب التشابه بين الصور. استخدمت هذه الاطروحة الكسور كخصائص محلية لتمثيل الصورة تستند الى مناطق بارزة في حين تبقى ثابتة لتغير نقطة النظر والاضاءة. تعتبر الكسور شائعه لقدرتها على استخراج ميزة التشابه الذاتي. ولفترة طويلة، استخدم الباحثون الكسور لضغط الصورة. على مدى السنوات الاخيرة، تم تطبيقها في التنقيب على البيانات. لهذه الاطروحة هدفين رئيسيين : اولا لدراسة القدرة على استخراج خصائص التشابه الذاتي من الصور دون استخدام بعد الكسور والذي يعتبر حساس للضوضاء العددية او التجريبية ومقيد بكمية البيانات. وثانيا لبناء الرسم البياني على اساس الميزات المستخرجة وتطور خوارزمية تجميع بالاعتماد على الرسم البياني.وينقسم النظام المقترح الى مرحلتين، بناء مصفوفة التشابه بواسطة طريقة كسورية وخوارزمية تنقيب المخططات. تم تطبيق PIFSلاستخراج ميزات التشابه الذاتي من صورة واحدة فقط. ولكن في هذه الدراسة كيفت PIFS لاستخراج ميزات التشابه الذاتي من العديد من الصور. بسبب ان PIFS تستغرق وقتا طويلا، فقد تم تكييفها للعمل مع تقنيات المطابقة والتقليل، وايضا تم استخدام الدالة الهاش للحد من تعقيد الوقت. واستخدم النظام المقترح مصفوفة تشابه لبناء المخطط ووضع خوارزمية عنقدة شبكية تعتمد على خصائص كسورية التوصيل بين العقد التي تمثل صور.استخدمت عدة بيانات لاختبار النظام المقترح. ولان النظام ينقسم الى مرحلتين، الاولى بناء مصفوفة التشابه والثانية هي خوارزمية تجميع الرسم البياني. لذلك، تم اختبار كل مرحلة بشكل منفصل. في الاول، يتم اختبار بناء مصفوفة التشابه (الميزات المستخرجة) مع خوارزمبة K - means لمعرفة صحة الميزات المستخرجة.وتم اقتراح طرائق لتقليل وقت التنفيذ ومقارنتها مع الطرائق التقليدية. وخفضت دالة الهاش التعقيد من O(m×n) الى O(m log⁡n) بينما قللت المطابقة والتقليل التعقيد الى O(m×n/t) حيث t عدد دوال المطابقة.اما طريقة التجميع البيانية المقترحة تم اختبار صحتها باستخدام البيانات الحقيقية واستخدمت المقاييس النمطية، الموصلية، التغطية، وكثافة الجودة وتم عرض النتائج والتحقق من صحتها من الناحية العددية والبصرية مع عدد عقد المختلفة. وقد اظهرت النتائج التي تم الحصول عليها دقة بين 0.80 و0.99 لجميع المقاييس.واظهرت النتائج ان للكسور قدرة كبيرة على استخراج ميزة التشابه الذاتي لاستخدامها في التنقيب عن الصور مثل التجميع. واعطت خصائص التشابه الذاتي كسورية نتائج جيدة. وان الميزات المستخرجة مشابه الى مصفوفة المجاورة التي يتم استخدامها لتمثيل الرسم البياني. لذلك، تعتبر بنية جيدة لتمثيل الرسم البياني. | Image clustering refers to the division of images into various sets of images. In this regard, each set known as cluster includes images that are similar in features to each other but different those of other sets. The global features as statistical features can be interpreted as a particular property of image involving all pixels were used to calculating similarity among images by most of the researchers. This thesis used fractal features as local features to represent an image based on salient regions while remaining invariant to viewpoint and illumination changes. Fractal is popular because of their ability to extract the self - similarity feature. For a long time, researchers used fractals for image compression. Over the latest years, they have been applied in mining. This thesis has two major purposes, first to studies the ability to extract fractal Self - similarity features from images without using fractal dimension which is sensitive to numerical or empirical noise and limitations in the amount of data. Second to constructs graph based on extracted features and develops graph cluster algorithm.The proposed system is divided into two phases, the Similarity Matrix construction by a fractal method and a Graph Clustering algorithm. Partitioned Iterated Function Systems (PIFS) is applied to extracting Self - similarity features from just one image. This study developed PIFS to extracting Self - similarity features from many of images. Since the PIFS algorithm is time - consuming, it has been adapted to work with Map - Reduce techniques and also hash function was used to reduce the time complexity. The proposed system used similarity matrix to construct a graph structure and developed a graph clustering algorithm based on connectivity fractal features among nodes that represents as images.Each phase was tested Separately. In the first phase, Similarity Matrix construction (features extraction) is tested with K - means clustering algorithm to find out the correct features extracted. The B - Cubed recall and precision are estimated with good results to precision and recall accuracy.Then proposed methods of reducing time complexity results is presented and compared with traditional methods. The hash function reduced the complexity O(m×n) to O(m log⁡n) while Map/reduce technique reduce the complexity O(m×n) to O(m×n/t) for time where t is a number a of map task.The second phase, Graph Clustering algorithm is tested with the real - world graph dataset. The clustering result was evaluated by Modularity, Conductance, Coverage, and Density Quality Metrics and the results were presented and validated both numerically and visually with different nodes number. The obtained results have shown accuracy between 0.80 and 0.99 for all metrics.

نظام الكشف التعاوني عن هجومات الفيضان الموزعة للحرمان من الخدمة والتعقب المستوحى من مجتمع العناكب الاجتماعية == Collaborative Detection System of DDoS Flooding Attacks and Tracing Inspired by Social Spiders Society

Author name: عادل محمد سلمان القريشي
Supervisor name: صفاء عبيس المعموري
General topic: Computer Science
Specific topic: Computer Science
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: لا تزال شبكة الانترنيت تعاني من المشاكل الامنية التي تهم بشكل رئيسي الاشخاص الذين يستخدمون اجهزتهم للاتصال بالانترنت، سواء كانوا افراد او مؤسسات كبيرة. الهجمات الموزعة للحرمان من الخدمة، لا تزال واحدة من اهم المواضيع التي يتم مناقشتها حاليا في تهديدات امن الشبكات للشركات التي تقدم الخدمات لعملائها. في هذه الاطروحة، تم اقتراح نظام الكشف التعاوني. واستند على مرحلتين : (1) مرحلة الكشف؛ (2) مرحلة التعقب. اعتمادا على الفكرة المستوحاة من مجتمع العناكب الاجتماعية، تم تصنيف اجهزة التوجيه الى نوعين، على النحو التالي : (1) جهاز التوجيه الذكر، الذي هو مرتبط مباشرة مع الخادم؛ (2) جهاز التوجيه الانثى، والذي هو كل جهاز توجيه غير مرتبط مباشرة مع الخادم. ويتميز النظام المقترح بانه حل قائم على جهاز التوجيه وعلى فحص التدفقات.يمكن تقسيم مرحلة الكشف الى اربع خطوات، على النحو التالي : (1) جمع البيانات؛ (2) معالجة البيانات واستخراج الميزات؛ (3) بناء نموذج التصنيف، باستخدام خوارزمية شجرة القرار عالية السرعة (VFDT) كخطوة للكشف المبكر، والتي سيتم استخدامها من قبل كل جهاز توجيه انثى في الشبكة؛ (4) كشف الشذوذ (الهجوم) باستخدام خوارزمية الغابات العشوائية (RF) للتصنيف، والتي سيتم تنفيذها في كل جهاز توجيه ذكر. الجمع بين هاتين الخوارزميتين سوف ينتج عنه خوارزمية تصنيف جديدة تسمى هوفدينغ الغابات العشوائية (HRF).تبدا مرحلة تتبع مصادر الهجوم عندما يتم العثور على بيانات الهجوم. جهاز التوجيه الذكر القريب من الخادم الضحية سوف يتتبع مصادر الهجوم بالاعتماد على قيمة الاهتزاز للتدفق، ثم رفع الانذار وارسال جميع المعلومات الى مسؤول الشبكة لاتخاذ الاجراءات اللازمة. وقد استلهمت قيمة الاهتزاز من مجتمع العناكب الاجتماعية، والذي هو قيمة تاثير جهاز التوجيه الانثى على كل تدفق يمر من خلاله.وقد تم استخدام برنامج محاكاة شبكة NS3 لتوليد بيانات الشبكة. ثم الحصول على النتائج واختبار النظام بواسطة برنامج مبرمج باستخدام لغة C++. وعلاوة على ذلك، طبقت عدة تجارب، وتم اعتماد تجربتين لاختبار النظام المقترح، الاول هو 90 ثانية، في حين ان الثانية هي 1200 ثانية. اجريت هذه التجارب لتوليد البيانات العادية وكذلك توليد بيانات هجوم الفيضان الموزع للحرمان من الخدمة للنوعين TCP وUDP. تم اختبار البيانات التي تم توليدها لاثبات ما اذا كانت مشابهة للبيانات الحقيقية عن طريق اختبار اثنين من الخصائص التي هي التباين العالي والتشابه الذاتي. وقد اظهرت النتائج ان البيانات التي تم توليدها لها نفس خصائص البيانات الحقيقية، وتمت الموافقة على نسبة حوالي 95٪.بالاضافة الى ذلك، لتقييم اداء خوارزمية هرف الجديدة، تم استخدام ثلاثة تدابير : (1) نسبة دقة التصنيف، والتي كانت 99.9983٪ و99.9990٪ على التوالي لكل من التجارب (90 ثانية و1200 ثانية). (2) معدل الكشف، والتي تبين 9.9996٪ و99.9997٪، على التوالي، لكلا التجربتين. و(3) نسبة الانذار كاذب، كان 0.016٪ و0.0088٪ على التوالي لكلا التجربتين. وكان متوسط وقت الكشف 21.71 و28.46 ثانية لكل من التجارب على التوالي.يستخدم النظام المقترح مبدا تقليل السمات المستخدمة في التصنيف، مما ادى الى انخفاض في حجم الذاكرة المستخدمة بنسبة 62.96٪ وانخفاض في مساحة القرص الثابت المستخدم بنسبة 51.75٪.واخيرا، في عملية البحث عن المفقودين، والوصول الى اقرب جهاز التوجيه الاناث الى مصدر الهجوم، حيث تم تحديد معظم هذه الموجهات، لكلا التجربتين، مع نسبة 100٪. | The Internet still suffers from security problems which are the main concern for those connected via their devices, whether they are individuals or institutions. The Distributed Denial of Service (DDoS) attacks are still one of the most significant current discussions regarding network security threats for companies providing services to their clients.In this dissertation, a collaborative detection system which proposed is based on two parts : (1) the Detection phase, and (2) the Tracing phase. Inspired by the social spider’s society, the routers were classified into two types : (1) Male router, which is near the server and directly connected with it; and (2) Female router, which is near the user and directly connected with it or between the user and the server. The proposed system is characterized as a router - based and flow - based solution.The detection phase can be divided into four steps : (1) data collection; (2) data preprocessing and extraction of features; (3) building the classification model, using a Very Fast Decision Tree (VFDT) algorithm as an early detection step, which will be used by each female router in the network; and (4) anomaly (attack) detection using the Random Forest (RF) algorithm for classification, which will be implemented in each male router. The combination of these two algorithms will generate a new classification algorithm called the Hoeffding Random Forest (HRF).The tracing phase will be started when the attack data is found. The male router near the victim server will trace the attack sources based on the value of the vibration of the flow, then raise the alarm and send all the information to the network administrator, to take an action. The vibration value has been inspired by the social spider’s society, which is the effect of the female router on each flow passing through it.NS3 network simulation software has been used to generate the network data. Then obtain the results and test the system by a software programmed by C++. Moreover, several experiments were applied, and two experiments were adopted to test the proposed system; the first is 90 seconds, while the second is 1200 seconds. These experiments were performed to generate normal data and DDoS flooding attack data for TCP and UDP types. The generated data has been tested to prove if it is similar to the real data by testing two critical characteristics : high - variability and self - similarity. The results show that the generated data has the same characteristics as the real data, and is approved with ratio approximately 95%.Additionally, to evaluate the performance of the new HRF algorithm, three measures have been used : (1) classification accuracy ratio, which was 99.9983% and 99.9990% respectively for both experiments (90 sec. and 1200 sec.); (2) detection rate, showing 9.9996% and 99.9997%, respectively, for both experiments; and (3) false alarm, was 0.016% and 0.0088% respectively for both experiments. The average of the detection time was 21.71 and 28.46 seconds for both experiments respectively.The proposed system uses the principle of reducing the features that used in the classification, which led to a reduction in the used memory size by 62.96% and a reduction in the used hard disk space by 51.75%.Finally, in the tracing process, accessing the nearest female router to the source of the attack, where most of these routers have been identified, for both experiments, with ratio 100%.

اخفاء المعلومات في صفحات الويب الهجينة == HYBRID STEGANOGRAPHY OF WEBPAGE

Author name: رسل عامر عبد المحسن
Supervisor name: وسام سمير بهية
General topic: Computer Science
Specific topic: Information Technology
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: اخفاء المعلومات هو علم اخفاء البيانات والمعلومات السرية في غطاء رقمي مثل ملفات نصيه او ملفات صوتيه او ملفات الفديوا او صوره او تاك بحيث يصعب على المشاهد العادي على معرفة وجود رسالة مخفية فيه. في هذه الرسالة تم اقتراح طرق جديده لاخفاء البيانات السرية والمهمة بالاعتماد على خوارزمية الاخفاء بالنص والصورة والتاك.ويتكون النظام المقترح من ثلاث مراحل. المرحلة الاول هي تحميل صفحه (webpage)، والمرحلة الثانية هي تحليل صفحة webpage))، نحلل الصفحة حتى نرى النسبة الاكثر التي تتكون منه الصفحة فيما اذا كان (نص او صوره او تاك)، والمرحلة الثالثة هي الاخفاء اما نخفي بالنص او الاخفاء بالنص او الاخفاء بالتاك. الاخفاء بالنص يتم من خلال تحويل الرساله السريه المراد اخفاءها الى (binary) وبعد ذلك نقوم باخفاء (bits) السريه في النص الموجود في صفحة الويب التي قمنا بتصميمها اذا كان الدينا ل(1) نكبر اول حرف من كل كلمة بدرجه واحده فقط بحيث تكون غير جالبه للشك اما اذا كان لدينا (0) نصغر اخر حرف من كل كلمة بدرجه واحده فقط من النص الموجود في صفحة الويب. اما الاخفاء بالصورة يتم من خلال تحويل الرسالة السري الى (binary) وبعد ذلك نقوم باخفاء (bits).اذا كان لدينا الbit(1) يتم اخفاءها في حال كان pixel)) في الاطار الاول والاطار الثاني زوجي اما اذا كان لدينا bit(0) يتم اخفاءها في حال كان pixel)) في الاطار الاول والاطار الثاني فردي.الاخفاء بالتاك يتم من خلال تحويل النص السري الى encoding) (وبعد ذلك نقوم باخفاء(bits) في التاك وهذه العملية تتم من خلال اذا كان (1) نجعل اول حرف من التاك كبير واذا كان (0) نجعل اول حرف من كلمة صغير. | Information hiding is a science of hiding data and secret information in digital cover such as (texts files, images or tags files), where difficult for the normal viewer to know the presence of a hidden message in it. This thesis suggests a method to hide the secret and important information based on Steganography.Sending encrypted messages frequently will draw the attention of third parties, i.e. crackers and hackers, perhaps causing attempts to break and reveal the original messages, in a digital world, Steganography is introduced to hide the existence of the communication by concealing a secret message inside another unsuspicious message.In this thesis, we have worked with data hiding techniques in (text, image and tags) of Internet webpages. In particular, we will propose an idea of using three Steganography techniques to hide secret information in (text, image and tags) of webpages. In this thesis, the webpage will be analyzed to check the size of different contents of webpages (text, images, and tags). Then, we will hide secret data in one or more contents of webpages depend on capacity of webpage contents and information which want to be hide. When hiding in the text by conversion the secret message to the binary, the hide secret bit (1) by enlarging size the first letter of each word to one degree only in cover text, either in case hide the secret bit (0) we work on reduce size the last letter of each word to one degree only in cover textIn addition, we propose a method of Steganography to hide text message in the image that is exist in the webpage through the conversion of secrete text to binary and then you will hide the bits in the frame of the image. In addition, we propose a Steganography in the Tags of the webpages, by encoding secrete bits and then hide these bits through tags cover webpages. Changing the letters of tags to uppercase letters and small case letters according to secrete bits. the propose technique characterized by highly capacity since we can hide more data when hiding in text, tags and images

تنقيب محتويات وبيانات استخدام الشبكة العنكبوتية بالاعتماد على تقنيات العنقدة المحدثة == Web Content and Usage Mining Based on Modified Clustering Techniques

Author name: احمد جبار عبيد
Supervisor name: توفيق عبد الخالق الاسدي
General topic: Computer Science
Specific topic: Computer Science
Degree: Doctorate
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: The extensibility of diversified information that available on the Web along with massive users' have accessed to the Web services frequently produce several challenges related to such critical tasks such as controlling, monitoring and perception of the Web contents. However, novel techniques must be used to satisfy the modernistic requirements and provides better understanding to the colossal collection of diversity data types that is growing in fast manner every day on the Web.Web Mining is an extension of Data Mining techniques upon the data that stored on the Web. Web Mining is classified into three categories based on the type of data that used in mining process which are : Web Content Mining (WCM) is concern with the process of extract useful information from Web pages' contents, Web Usage Mining (WUM) is concern with discovering users' access pattern from Web usage data, and finally Web Structure Mining (WSM) is concern with extracting knowledge from the structure of the hyperlinks. Web documents are the most complex data that scattered on the Web in random way and a lot of these documents are created without any prior information. Unsupervised Data Mining Clustering technique, is one of the most usage techniques that aim to portioned out the objects into set of coherence groups, where the objects in a cluster are having common patterns than objects in other clusters.In this dissertation, the task of Web Mining is divided into two parts based on the data collected from the universities of (Kufa, Technology, Anbar and Diyala). First part is hold the Web documents by applying WCM techniques upon the Web Pages and Images of the universities Web sites, while second part is consider applying WUM techniques upon the Web usage data that collected from the Kufa university Web server. Proposed system consist of two parts : first part uses a novel approach to pre - process and extract unobserved patterns from Web pages' text blocks content,

محاكاة لتقيم اداء شبكة المركبات اللاسلكية المخصصة (VANET) == Simulation of Performance Evaluation For VANETs

Author name: زهراء ياسين حسن
Supervisor name: سعد طالب حسون الجبوري
General topic: Computer Science
Specific topic: Computer Science
Degree: Master
University: University of Babylon - Information Technology Collage - Department Of Software
Language: English
University location: Babylon
First pages:
Abstract: النسخة الجديدة المعدلة من شبكات العقدالمتنقلة تدعى شبكات المركبات المخصصة والتي تعتبر واحدة من مشاريع "انترنت الاشياء" . يهدف VANET لتوظيف واستخدام تقنيات مخصصة لتخفيف الازدحام على الطرق في الوقت الحقيقي. لديها تاثير كبير على المجتمع عن طريق الحد من الوقت الذي يستغرقه السفر، واستهلاك الوقود، حياة الركاب واخيرا لتوفير المال. في بيئة شبكات المركبات المخصصة تمثل كلمة مركبة عقدة ذكية مع القدرة على التواصل مع الجيران المتحركة في الشبكة. VANET يقدم جوانب من التحديات اكثر مقارنة بال (MANET) بسبب كثرة تنقل العقد وتغييرات الهيكل سريعة في VANET. وقد تم تصميم عدة بروتوكولات التوجيه قدمها الباحثون بعد النظر في التحديات الكبرى المشاركة في VANETs. في هذه الاطروحة عدة منهجيات مقترحة وهي(نهج البث , نهج العنقدة, نهج محرك الجدول ونهج الاستعلام) لغة Net Logo هي اداة المحاكاة المقترحة لاستخدامها في برمجة,تصميم , خلق وتطبيق عدة هياكل لمختلف خرائط الطرق وهي (الطريق المفرد, الطريق المزدوج , طريق التقاطعات وطريق الاستعلام) . بيئة الطريق وغيرها من العوامل واللوائح واقتراحات الحالة سوف تختلف عن الاخرين. ثلاثة اشكال جديدة من نماذج الطرق هي مقترحة لتكون مطبقة ومجربة بواسطة نظام VANET . المحاكاة سوف تكون مستخدمة لملاحظة وقياس السلوكيات بواسطة توليد وجمع البيانات المطلوبة والعوامل والمعاملات. عدة مقاييس اداء يمكن ان تستخدم لتقيم اداء الشبكة مثل PDF,NRL , الانتاجية, عدد الرسائل المستلمة وعدد الرسائل المفقودة . النتائج النهائية اوضحت بان عدد المركبات, سرعة المركبات , ازدحام الطرق ومدى التغطية تمتلك تاثير ملموس على اداء شبكات ال VANET . | The new modified version of mobile Ad hoc Network called "Vehicular Ad hoc Network" (VANET) which considered as one of the projects of Internet of Things (IoT). VANET aims to employ and utilize ad hoc technologies to alleviate Road Congestion in real time. It has a great impact on the society by reducing travel time, fuel consumption, passenger's life and finally to save money. In VANET's environment, the word Vehicle represents an intelligent node with a capability to communicate with mobile neighbors in the network. VANET introduces more challenges aspects as compare to Mobile Ad - Hoc Network (MANET) because of high mobility of nodes and fast topology changes in VANET. Several routing protocols have been designed and presented by researchers after considering the major challenges involved in VANETs. In this thesis, several approaches : (Broadcast approach, Clustering approach, Table - Driven approach and Queering road approach) are suggested, Net Logo simulator is proposed to be used in programming, designing, creating and implementing several structures for different road maps. The road environment and other factors, regulations and case suggestions will differ from others. Foure new road model forms : ( Single way road model, Two way road model, Junction road model and Queering road model) are suggested to be implemented and tested by VANET system. The simulation will be used to observe and measure their behaviors by generating and collecting the required data, factors and parameters. Many performance metrics can also be used to evaluate this network behavior; such as Packet Delivery Fraction (PDF), Throughput, End to end delay, dropped and received messages. The final results showed that the number of Vehicles, Vehicles speeds, traffic intensity and the coverage area having a tangible effect on the performance of VANET’s networks.
1 ... 3 4 5 6