Thesis Title: Modified “Detection and Recognition” Algorithm for Arabic Text Using Convolutional Neural Networks
Author: Oday Salim Albayati, Supervisor: Dr. Mouyad A. Fadhil Al-Athami, Year: 2020
Faculty: Information Technology, Department: Computer Science

Abstract: Extracting textual information from images is one of the important challenges in computer science and a difficult issue with numerous useful applications like self-driving cars and translation. Arabic text detection and recognition in natural images have received increasing attention in computer vision and image understanding because the Arabic text contains many font styles that make recognition a challenging task. In this research we used a new algorithm for Arabic text detection and recognition based on machine learning by using YOLOv3 (You Only Look Once) is a Convolutional Neural Network (CNN) and region-based text detection approach. Started from a review of related works in Arabic text, collected our dataset by taken images contain Arabic text for sign area in Jordan, prepared the dataset for training by labeling each word in these images, illustrates position Arabic text in each image in the dataset by using bounding box technique. Then the training process is done by using the Darknet-53 framework that builds on YOLOv3 architecture. Also, testing our model and evaluating the result with previse work. In this research, we used 218 classes to represent the 218 words in Arabic language and archive up to 90?curacy rate and high speed less than 0.4 milliseconds on GPU for detection and recognition tasks.

Keywords: Arabic text, YOLOv3, detection, recognition, CNN

Thesis Title: Modified classification for brain tumors in Magnetic Resonance Imaging (MRI) using region convolutional neural networks (R-CNN)
Author: Abdelqader Waleed Al-Gailany, Supervisor: Dr. Mouyad A. Fadhil Al-Athami, Year: 2020
Faculty: Information Technology, Department: Computer Science

Abstract: Brain tumor detection and recognition in machine learning is a trendy challenge in computer science. Cause of difficult procedure also decide the type of the tumor (glioma, meningioma, pituitary ...etc.). Brain tumor is one of the most dangerous diseases that happens to the human body. Magnetic resonance imaging is a way to imaging the brain to assist the radiologist. This research focus on recognize the normal and abnormal brains. Also, detect and recognize tumor type in abnormal brains. This goal will be achieved by artificial intelligence using deep learning with Mask Region convolutional Neural Networks. This is a trendy way to classify and detect the objects with less time in training. This framework faster and accurate than the others with multiclass classification. This research achieves in average 95?curacy and results in detection and recognition. At last this research can detect and recognize the tumor and its types.

Keywords: Artificial intelligence, Deep learning, CNN, R-CNN, Mask R-CNN, Brain tumor

Author: Khaled Jehad Zakariya, Supervisor: Dr. Tarek A. Tutunji, Year: 2020
Faculty: Engineering, Department: Mechatronics Engineering

Abstract: This Thesis investigates the modelling, trajectory planning, and control simulation of 6-DOF KUKA KR6-2 robot manipulator. The study performed uses theoretical analysis, mathematical modelling, and Matlab simulation. The work is divided into sequential stages that include: kinematic modelling, trajectory planning, robot dynamics simulation, and control tuning. In all these stages, Matlab models for the KUKA KR6-2 are developed. Additionally, the equations for forward and inverse kinematics are utilized. These equations are used to move the robot between the joint-space and the workspace. An imported CAD model of KR6-2 parts and links are used in the developed Matlab model for 3D visual simulation. Trajectory planning methods, using Splines, are applied in order to move the manipulator from an initial pose to the final pose. Transformation matrices are used to map between the joint space and the workspace so that the end-effector to be moved smoothly and passed all subject points in the desired time. These analyses are important for robot manipulators in order to execute specific task such as following a preplanned path. The practical geometric parameters, physical characteristics, and restriction relations are adopted to establish robots dynamics and kinematics model. The SimMechanics toolbox is used to build the dynamic model of the KR6-2 in Matlab and a proportional-integral-derivative (PID) controller is applied at the joint level. Simulation results show that the developed Matlab models succeed in modelling the robot kinematic, planning its trajectory, and controlling the position and velocity of the manipulator.

Keywords: trajectory, simulation, manipulator

Thesis Title: The effect of internal audit on the quality of accounting information in the Jordanian public shareholding companies
Author: Louna Katbeh Bader, Supervisor: , Year: 2020
Faculty: Business, Department: Accounting

Abstract: This study aimed to identify the effect of internal audits on the quality of accounting information in the Jordanian public shareholding companies. Population is made up of a group of employees in the accounting and internal auditing department at Jordanian public shareholding companies, while the study sample size is (320) respondents consist of internal auditors, financial managers, accountants, employees in the internal audit department, and members of audit committees in Jordanian public shareholding companies. The researcher used a questionnaire as a survey method in addition to multiple regressions in the hypothesis test. The main survey method in addition to the study concluded that there is a significant effect of internal audit on the quality of accounting information in Jordanian public shareholding companies that is statistically proven through four dimensions: (Independence of internal audit, the efficiency of internal audit, professional care, commitment to performance standards), and the internal audit effect was found Positive Medium between all the independent and dependent variables of the study.The results of the study confirmed that the internal auditors have independence, efficiency, professional care, and adherence to performance standards, which make their contribution effective toward improving the quality of accounting information. This indicates that the greater the auditors' commitment to the independent variables, the greater the role of the internal audit in improving the quality of accounting information.One of the most important recommendations of the study is to actively use its results in compliance with the national interest and the Jordanian economy, by improving accounting information to reduce negative impacts which lead to the reluctance of foreign investments on stakeholders and the Jordanian economy as a whole, reduce poverty and unemployment, and debt burdens. The second recommendation is the significance of the continuous development of all laws, legislations, and standards related to internal auditing to achieve better quality, which improves the accounting information.

Keywords: Internal Audit, quality of accounting information, Jordanian shareholding companies

Thesis Title: جماليات المكان في تجربة جلال برجس الروائية
Author: وائل أحمد فتحي الزغايبة, Supervisor: الدكتور غسان عبد الخالق, Year: 2020
Faculty: Arts, Department: Arabic Language and Literature

Abstract: This study stood at the aesthetics of the place in the experience of Jalal Burgess, the novelist, through research, analysis, tracking, deduction, and induction, in three of his novels, which are: the novel (Guillotine Dreamer), the novel (Fire Snakes), and the novel (The Five Senses Ladies). The study monitored the types of places in each novel, and divided them into: (closed places, open places, and imagined places), and tracked the specificity of these places, the aesthetic nature of their use in each novel in a way that suits their narrative context, its thematic format, and its artistic structure that affected the places by doing The integrative relationship with the rest of the novel, which the novelist did not overlook.

Keywords: aesthetic, interpretative, and pilgrimage function