Faculty Of Computer Science Research Paper

Permanent URI for this collectionhttp://185.252.233.37:4000/handle/123456789/304

Browse

Now showing 1 - 20 of 188

Hyperparameters Optimization of Deep Convolutional Neural Network for Detecting COVID-19 Using Differential Evolution
(Springer, 01/01/2022) Ezzeldin Nagib, Abdelrahman; Saeed, Mostafa Mohamed; El-Feky, Shereen Fathy; Khater Mohamed, Ali
COVID-19 is one of the most dangerous diseases that appeared during the past 100 years, that caused millions of deaths worldwide. It caused hundreds of billions of losses worldwide as a result of complete business paralysis. This reason has attracted many researchers to attempt to find a suitable treatment for this dreaded virus. The search for a cure is still ongoing, but many researchers around the world have begun to search for the safest ways to detect if a person carries the virus or not. Many researchers resorted to artificial intelligence and machine learning techniques in order to detect whether a person is carrying the virus or not. However, many problems are arising when using these techniques, the most important problem is the optimal selection of the parameter values for these methods, as the choice of these values greatly affects the expected results. In this chapter, Differential Evolution algorithm is used to determine the optimal values for the hyperparameters of Convolutional Neural Networks, as Differential Evolution is one of the most efficient optimization algorithms in the last two decades. The results obtained showed that the use of Differential Evolution in optimizing the hyperparameters of the Convolutional Neural Network was very efficient.
YOLO fish detection with Euclidean tracking in fish farms
(Springer, 01/03/2021) Wageeh, Youssef ; Mohamed, Hussam El‑Din ; Fadl, Ali ; Anas, Omar ; ElMasry, Noha ; Nabil, Ayman ; Atia, Ayman
The activities of managing fish farms, like fish ponds surveillance , are one of the tough and costly fish farmers’ missions. Generally, these activities are done manually, wasting time and money for fish farmers. A method is introduced in this paper which improves fish detection and fish trajectories where the water conditions is challenging. Image Enhancement algorithm is used at first to improve unclear images. Object Detection algorithm is then used on the enhanced images to detect fish. In the end, features like fish count and trajectories are extracted from the coordinates of the detected objects. Our method aims for better fish tracking and detection over fish ponds in fish farms.
Automatic Identification of Student’s Cognitive Style from Online Laboratory Experimentation using Machine Learning Techniques
(Institute of Electrical and Electronics Engineers, 04/12/2021) Yousef, Ahmed Mohamed Fahmy; Atia, Ayman; Youssef, Amira; Saad Eldien, Noha A; Hamdy, Alaa; Abd El-Haleem, Ahmed M; Elmesalawy, Mahmoud M
Online learning has emerged as powerful learning methods for the transformation from traditional education to open learning through smart learning platforms due to Covid-19 pandemic. Despite its effectiveness, many studies have indicated the necessity of linking online learning methods with the cognitive learning styles of students. The level of students always improves if the teaching methods and educational interventions are appropriate to the cognitive style of each student individually. Currently, psychological measures are used to assess students' cognitive styles, but about the application in virtual environment, the matter becomes complicated. The main goal of this study is to provide an efficient solution based on machine learning techniques to automatically identify the students' cognitive styles by analyzing their mouse interaction behaviors while carrying out online laboratory experiments. This will help in the design of an effective online laboratory experimentation system that is able to individualize the experiment instructions and feedback according to the identified cognitive style of each student. The results reveal that the KNN and SVM classifiers have a good accuracy in predicting most cognitive learning styles. In comparison to KNN, the enlarged studies ensemble the KNN, linear regression, neural network, and SVM reveal a 13% increase in overall total RMS error. We believe that this finding will enable educators and policy makers to predict distinct cognitive types in the assessment of students when they interact with online experiments. We believe that integrating deep learning algorithms with a greater emphasis on mouse location traces will improve the accuracy of our classifiers' predictions. © 2021 IEEE.
Applying Deep Learning to Track Food Consumption and Human Activity for Non-intrusive Blood Glucose Monitoring
(Institute of Electrical and Electronics Engineers, 04/12/2021) Samir, Mohamed Amr; Mohamed, Zeinab A; Hussein, Mona Abdelmotaleb A; Atia, Ayman
Blood glucose monitoring is a wide area of research as it plays a huge part in controlling diabetes and many of its symptoms. A common human disease 'Diabetes Mellitus' (DM), which is characterized by hyperglycemia, has a number of harmful complications. In addition, the low glucose level in blood caused by hypoglycemia is correlated to fatal brain failure and death. In this paper, we explore a variety of related research to have a grasp on some of the systems and concepts that can assist in forming an autonomous system for glucose monitoring, including deep learning techniques. The proposed system in this paper utilizes non-intrusive Continuous Glucose Monitoring (CGM) devices for tracking glucose levels, combined with food classification and Human Activity Recognition (HAR) using deep learning. We relate the preprandial and peak postprandial glucose levels extracted from CGM with the Glycimc Load (GL) present in food, which makes it possible to form an estimation of blood sugar increase as well as predict hyperglycemia. The system also relates human activity with decrease in blood glucose to warn against possible signs of hypoglycemia before it occurs. We have conducted 3 different experiments; two of which are comparison between deep learning models for food classification and HAR with good results achieved, as well as an experimental result that we obtained by testing hyperglycemia prediction on real data of diabetic patients. The system was able to predict hyperglycemia with an accuracy percentage of 93.2%. © 2021 IEEE.
A NOVEL OVERSAMPLING TECHNIQUE TO HANDLE IMBALANCED DATASETS
(European Council for Modelling and Simulation, 06/01/2020) Mahmoud, A; Ali, F; El-Kilany, A; Mazen, S
With the amount of data is growing extensively in different domains in the recent years, the data imbalance problem arises frequently. A dataset is called imbalanced when the data of a certain class has significantly more instances than that of other classes of the same dataset. This imbalanced nature of the data negatively affects the performance of a classifier since misclassification of data may cause data analysis results to be inaccurate and hence leads to wrong business decisions. This paper presents a study of the different techniques that are used to handle the imbalanced dataset, and finally proposes a novel oversampling technique to tackle the binary classification of imbalanced dataset problem. © ECMS Mike Steglich, Christian Mueller, Gaby Neumann, Mathias Walther (Editors).
Examining the Effects of Embodied Interaction Modalities on Students' Retention Skills in a Real Classroom Context
(SPRINGER HEIDELBERG, TIERGARTENSTRASSE 17, D-69121 HEIDELBERG, GERMANY, 06/01/2022) Chettaoui, Neila ; Atia, Ayman ; Bouhlel, Med. Salim
Embodied cognition theory denotes that knowledge is incorporated into the body's sensorimotor system, which facilitates learning and understanding abstract concepts. In this context, several interaction modalities have been introduced to design learning experiences that promote multisensory processing. This study examined the impacts of the type of embodied interaction modality on learning gains in a real classroom context. The researchers designed learning interfaces involving different interaction modalities including tablet application, tangible user interface, motion-based technology, and multimodal interaction. Thirty-six primary school students (aged 7 to 9) were assigned to four groups to learn the basics of the human body anatomy. The study adopted an immediate and a 20-day delayed post-test to measure students' knowledge retention. Regardless of interaction modality type, participants showed significant immediate learning gains. However, participants in the multimodal embodiment conditions performed better on the delayed post-test. The findings suggested that multimodal embodied interaction, merging between body movements and tangible user interfaces, may lead to better knowledge retention. The process of performing body movements and physical interaction offered an alternative and a complementary encoding strategy for understanding and memorizing the learning concepts.
TGT: A Novel Adversarial Guided Oversampling Technique for Handling Imbalanced Datasets
(Elsevier, 1/19/2021) Mahmoud, Ayat; El-Kilany, Ayman; Ali, Farid; Mazen, Sherif
With the volume of data increasing exponentially, there is a growing interest in helping people to benefit from their data regardless of its poor quality. One of the major data quality problems is the imbalanced distribution of different categories existing in the data. Such problem would affect the performance of any possible of analysis and mining on the data. For instance, data with an imbalanced distribution has a negative effect on the performance achieved by most traditional classification techniques. This paper proposes TGT (Train Generate Test), a novel oversampling technique for handling imbalanced datasets problem. Using different learning strategies, TGT guarantees that the generated synthetic samples reside in minority regions. TGT showed a high improvement in performance of different classification techniques when was experimented with five imbalanced datasets of different types. 2021 Production and hosting by Elsevier B.V. on behalf of Faculty of Computers and Information, Cairo University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
An integrated classification method for brain computer interface system
(IEEE, 10/07/2015) Mousa, Farid A.; El-Khoribi, Reda A.; Shoman, Mahmoud E.
A channel of communication for both human brain and computer system is provided via a system called Brain Computer Interface (BCI). The vital aim of BCI research is to develop a system that helps the disabled people to interact with other persons and allows their interaction with the external environments or as an additional man-machine interaction channel for healthy users. Different techniques have been developed in the literature for the classification of brain signals. The purpose of this work is to deveolp a novel method of analyzing the EEG signals. We have used high pass filter to remove artifacts, DWT algorithms for feature extraction and features like Mean Absolute Value, Root Mean Square, and Simple Square Integral are used. The neural network algorithm is used to find the correct class label for EEG signal after clustering the feature vectors using K-Nearest Neighbor algorithm. It has been depicted from results that the proposed integrated technique outperforms a better performance than methods mentioned in literature
Generalized Adaptive Differential Evolution algorithm for Solving CEC 2020 Benchmark Problems
(Institute of Electrical and Electronics Engineers Inc., 10/26/2020) Mohamed, A.K; Hadi, A.A; Mohamed, A.W
The effort devoted in introducing new numerical optimization benchmarks has attracted the attention to develop new optimization algorithms to solve them. Very recently, a new suite on bound constrained optimization problems is proposed as a new addition to CEC benchmark series. Differential Evolution (DE) is a simple Evolutionary Algorithm (EA) which shows superior performance to solve many CEC benchmark during the past years. This paper presents a new extension to DE algorithm through extending the line of research for AGDE algorithm. The new algorithm, which we name GADE, enhanced the DE algorithm by introducing a generalized adaptive framework for enhancing the performance of DE. Numerical experiments on a set of 10 test problems from the CEC2020 benchmarks for 5, 10, 15 and 20 dimensions, including a comparison with state-of-the-art algorithm are executed. Comparative analysis indicates that GADE is superior to other state-of-the-art algorithms in terms of stability, robustness, and quality of solution. © 2020 IEEE.
ISwimCoach: A Smart Coach guiding System for Assisting Swimmers Free Style Strokes
(MeC'20 Workshop, 10/29/2020) Ehab, Mohamed; Ahmed, Hossam Mohamed Mohamed; Hammad, Mostafa; ElMasry, Noha; Atia, Ayman
In sports, coaching remains an essential aspect of the efficiency of the athlete’s performance. This paper proposes a wrist wearable assistant for the swimmer called iSwimCoach. The key aim behind the system is to detect and analyze incorrect swimming patterns in a free crawl swimming style using an accelerometer sensor. iSwimCoach collects patterns of a swimmer’s stream which enables it to detect the strokes to be analyzed in real-time. Therefore, introducing quick and efficient self-coaching feature for mid-level athlete to enhance their swimming style. In our research, we were able to monitor athlete strokes underwater and hence assist swimming coaches. The proposed system was able to classify four types of strokes done by mid-level players (correct strokes, wrong recovery, wrong hand entry and wrong high elbow). The system informs both the swimmer and the coach when an incorrect movement is detected. iSwimCoach achieved 91% accuracy for the detection and classification of incorrect strokes by a fast non expensive dynamic time warping algorithm. These readings analyzed in real-time to automatically generate reports for the swimmer and coach.
A Large Dataset Enhanced Watermarking Service for Cloud Environments
(Springer, Cham, 11/28/2014) Zawawi, Nour; Hamdy, Mohamed; El-Gohary, Rania; Fahmy Tolba, Mohamed
Preventing data abuses in cloud remains an essential point of the research. Proving the integrity and non-repudiation for large datasets over the cloud has an increasing attention of database community. Having security services based on watermarking techniques that enable permanent preservation for data tuples in terms of integrity and recovery for cloud environments presents the milestone of establishing trust between the data owners and the database cloud services. In this paper, an enhanced secure database service for Cloud environments (EWRDN) is proposed. It based over enhancements on WRDN as a data watermarking approach. The proposed service guarantees data integrity, privacy, and non-repudiation recovering data to its origin. Moreover, it gives data owner more controlling capabilities for their data by enabling tracing users’ activities. Two compression categories to recover data to its origin introduced for the proposed service. Two compression technique (the arithmetic encoding and the transform encoding) chosen to represent each type. For large data sets, it has been proven that, the arithmetic encoding has a fixed recovery ratio equal to one. At the same time, the transform encoding saves space and consumed less time to recover data. Moreover, testing the performance is done of the proposed service versus a large number of tuples, large data set. The performance quantified in terms of processing time and the required memory resources. The enhanced EWRDN service has shown a good performance in our experiments.
A Novel Watermarking Approach for Data Integrity and Non-repudiation in Rational Databases
(Springer, Berlin, Heidelberg, 12/08/2012) Zawawi, Nour; El-Gohary, Rania; Hamdy, Mohamed; Fahmy Tolba, Mohamed
Keepinglarge scale data sets like data warehouse seems a vital demand of business organizations. Proving copyright, ownership, integrity and non-repudiation have a growing interest of database community. Many watermarking techniques have been proposed in the literature to address these purposes. This paper introduces a new technique WRDN (Watermarking Rational Database with Non-Repudiation) to protect the ownership of relational database by adding only one hidden column with a secret formula where it has the ability to know the latest updates made by users. The calculation of this formula is based on the values of other numeric and textual columns. Moreover, the proposed approach keeps track on the latest updates made to numeric and textual data by users. This approach is compared to two other alternative approaches. The proposed approach survives by 100% against insertion and deletion attacks.
Exploring the Impact of Multimodal Adaptive Learning with Tangible Interaction on Learning Motivation
(Institute of Electrical and Electronics Engineers Inc., 12/15/2020) Chettaoui, N; Atia, A; Bouhlel, M.S
Embodied learning defines a contemporary pedagogical theory focusing on ensuring an interactive learning experience through full-body movement. Within this pedagogy, several studies in Human-Computer Interaction have been conducted, incorporating gestures, and physical interaction in different learning fields. This paper presents the design of a multimodal and adaptive space for embodied learning. The main aim is to give students the possibility to use gestures, body movement, and tangible interaction while interacting with adaptive learning content projected on the wall and the floor. Thus, this study aims to explore how tangible interaction, as a form of implementing embodied learning, can impact the motivation of students to learn compared to tablet-based learning. Eighteen primary school students aged nine and ten years old participated in the study. The average percentages of answers on the Questionnaire on Current Motivation (QCM) pointed out a higher motivation among students learning via tangible objects. Results revealed a positive score for the Interest of learning abstract concepts using a tangible approach with a mean score of 4.78, compared to 3.77 while learning via a tablet. Furthermore, Success and Challenge measures, with a mean score of 4.67 and 4.56 indicate that physical interaction via tangible objects leads to significantly higher motivation outcomes. These findings suggest that learning might benefit more from a multimodal and tangible physical interaction approach than the traditional tablet-based learning process. © 2020 IEEE
Priority based Fuzzy Decision Multi-RAT Scheduling Algorithm in Heterogeneous Wireless Networks
(INT JOURNAL COMPUTER SCIENCE & NETWORK SECURITY-IJCSNS,, 12/30/2019) Hamda Othman, S; Asklany, S; Ali Mansouri, W
In this paper, we focus on the scheduling algorithm of heterogeneous wireless networks where the traffic has different requirements constrains which are needed to be fulfilled without exceeding the internal constrains of the node. We propose a priority based Fuzzy decision packet scheduling algorithm which is mainly based on delay, channel quality conditions, type of call (Handoff or new call) and classes of service. Using those parameters as an input, the Fuzzy logic approximates decision making using natural language terms instead of quantitative terms. The inputs to the Fuzzy system are fuzzified, implicated, aggregated and defuzzified to obtain the crisp value which represents the packet priority index. The simulation of the presented method is performed using Matlab and the results show that the presented scheme satisfactorily performs the system requirements and they prove that inclusion of the Fuzzy approach on the scheduler improves the packet drop ratio and minimize the end-to-end delay.
Parallelization of One Dimensional First Fit Decreasing Algorithm
(Institute of Electrical and Electronics Engineers Inc., 16/12/2021) Wessa, Eriny; Atia, Ayman
Bin packing is an optimization problem defined as placing different sized objects into similar containers or bins to minimize the number of used bins. This problem has different variations based on the dimensions of the bins, placement limitations, and priority. This paper focuses on one-dimensional bin packing. Two algorithms are explored in this paper, which are First Fit, First Fit Decreasing. The contribution of the paper is to explore the effect of parallelization on the First Fit Decreasing algorithm regarding the processing time and the utilization rate. Furthermore, the effect of using different numbers of concurrent workers on the problem is also explored. We proved that just by parallelizing the sorting in the First Fit Decreasing algorithm, we can decrease the computation time by 4.73% without affecting the utilization rate. © 2021 IEEE.
Exam Cheating Detection System with Multiple-Human Pose Estimation
(Institute of Electrical and Electronics Engineers, 19/11/2021) Samir, Mohamed Amr; Maged, Youssef; Atia, Ayman
Cheating in exams is a persistent problem that contributes to academic dishonesty. In this paper we explore a variety of related work proposed as a solution for exam cheating, then we propose an exam cheating detection system that works for both on-site and online examinations. The proposed system applies Human Pose Estimation that includes both single-user and multiple-user tracking algorithms. Based on video footage, the system can detect whether or not a student is cheating by continuously validating their head posture and hand movement conditions during the exam. The system doesn't fully imply a student is cheating, instead, we use the term 'warning' for the output to indicate that the student has met an abnormal condition that is similar to cheating behavior. At last, we validate the system usage in real-life examination environments through two different experiments that resulted in accuracy numbers of 92%-97% in cheating detection.
Untitled
(2001) Ali, Khayri A.M; CARLSSON, MATS; Pontelli, Enrico; GUPTA, GOPAL; HERMENEGILDO, MANUEL V.
Since the early days of logic programming, researchers in the field realized the potential for exploitation of parallelism present in the execution of logic programs. Their high-level nature, the presence of nondeterminism, and their referential transparency, among other characteristics, make logic programs interesting candidates for obtaining speedups through parallel execution. At the same time, the fact that the typical applications of logic programming frequently involve irregular computations, make heavy use of dynamic data structures with logical variables, and involve search and speculation, makes the techniques used in the corresponding parallelizing compilers and run-time systems potentially interesting even outside the field. The objective of this article is to provide a comprehensive survey of the issues arising in parallel execution of logic programming languages along with the most relevant approaches explored to date in the field. Focus is mostly given to the challenges emerging from the parallel execution of Prolog programs. The article describes the major techniques used for shared memory implementation of Or-parallelism, And-parallelism, and combinations of the two. We also explore some related issues, such as memory management, compile-time analysis, and execution visualization.
A novel ferrite cross-patch antenna for RCS reduction
(INT INST INFORMATICS & SYSTEMICS, 2002) Elshafiey, TF; Aberle, JT
This paper presents an efficient approach for the evaluation of the RCS of cross- patch antennas on ferrite substrates. The direction of the do magnetization of the ferrite is assumed arbitrarily in theta and phi. The analysis is based on a 3-D full wave analysis using method of moments. The Greens function and the excitation vector are formulated in a closed form using the transmission matrix approach. The linearly-varying basis function is used. A novel cross patch antenna is described and its performance compared with the full patch is also presented. The cross patch antenna minimizes number of RCS resonances and RCS levels.
Parallel Generational Copying Garbage Collection Schemes for Shared-Memory Multiprocessors
(The Military Technical College, 2003) A. M. Ali, Khayri; A. Omara, Fatma; A. Elshakankiry, Osama
In this paper, an improved parallel generational copying real-time garbage collection scheme for shared-memory multiprocessors, which supports load balancing among workers, has been proposed, implemented, and evaluated. The basic idea of improvement is developed from Ali's two papers [2,3]. The scheme proposed here is a form of copying collectors that attempt to eliminate the drawback of frequently copying long-lived (stable) objects. This class of schemes is called generational-based schemes, which is based on concentrating the collection efforts on small areas of memory, so-called young generation. They reduce the need for collecting the remaining large areas of memory, old generation. A modified scheme, without real-time response, has also been implemented and compared to the real-time one. A comparative study has been done for the two schemes with other two parallel non-generational copying garbage collection schemes founded by the author in [1]. According to this comparative study, we proved that the performance of generational schemes is better than the performance of non-generational schemes. Also, the overheads occurred due to real-time response have been calculated.
Object-Oriented Design Quality Models A Survey and Comparison
(2004) El-Wakil, Mohamed; El-Bastawisi, Ali; Boshra, Mokhtar; Fahmy, Ali
Since 1994, many Object-Oriented Design (OOD) quality models had appeared. OOD quality models aim is assessing OOD quality characteristics, such as maintainability, in a quantitative way through establishing relationships between OOD quality characteristics, and metrics computable from OOD diagrams, such as Depth of Inheritance Tree (DIT). This paper presents the results of our survey of the major OOD quality models appeared in literature since the MOOSE model in 1994, till the QMOOD model in 2002, then it proposes a set of desirable properties that should be possessed by OOD quality models and acomparison among the presented models with respect to the proposed desirable properties set

Browse

Browsing Faculty Of Computer Science Research Paper by Issue Date