AUS Repository (American University of Sharjah)
Not a member yet
2669 research outputs found
Sort by
A Hybrid Transfer Learning Approach to Teeth Diagnosis Using Orthopantomogram Radiographs
No full textThe rise in the emphasis on oral diseases has elevated the need to automate the diagnostic process of such diseases. Fortunately, the availability of modern computing devices has made the automated diagnosis of teeth readily possible using deep learning. Despite this, concerns about the accuracy and function of automated diagnosis remain among patients. To showcase the performance of such algorithms, we propose two approaches for the task of teeth diagnosis utilizing Orthopantomograms (panoramic radiographs): 1) a direct classification approach; and 2) a hybrid approach that combines a deep learning model with a traditional classifier. The results revealed that all ten chosen deep learning models experienced a similar or improved performance when used in conjunction with a machine learning classifier. In particular, Vision Transformer (ViT) performed the best with a record accuracy of 96% using both the direct and hybrid approaches. However, the hybrid framework combining AlexNet with a Support Vector Machine achieved an accuracy of 94%, and although it falls short of ViT in terms of performance, it comprises far fewer parameters. This highlights the approach’s effectiveness in improving performance without the need to use a deeper model, making it well-suited for clinical adoption where efficiency is important.Open Access Program from the American University of Sharja
Scaffold‑based 3D cell culture models in cancer research
No full textThree-dimensional (3D) cell cultures have emerged as valuable tools in cancer research, offering significant advantages over traditional two-dimensional (2D) cell culture systems. In 3D cell cultures, cancer cells are grown in an environment that more closely mimics the 3D architecture and complexity of in vivo tumors. This approach has revolutionized cancer research by providing a more accurate representation of the tumor microenvironment (TME) and enabling the study of tumor behavior and response to therapies in a more physiologically relevant context. One of the key benefits of 3D cell culture in cancer research is the ability to recapitulate the complex interactions between cancer cells and their surrounding stroma. Tumors consist not only of cancer cells but also various other cell types, including stromal cells, immune cells, and blood vessels. These models bridge traditional 2D cell cultures and animal models, offering a cost-effective, scalable, and ethical alternative for preclinical research. As the field advances, 3D cell cultures are poised to play a pivotal role in understanding cancer biology and accelerating the development of effective anticancer therapies. This review article highlights the key advantages of 3D cell cultures, progress in the most common scaffold-based culturing techniques, pertinent literature on their applications in cancer research, and the ongoing challenges.American University of SharjahAl-Jalila FoundationAl Qasimi FoundationPatient’s Friends Committee-SharjahBiosciences and Bioengineering Research InstituteGCC Co-Fund ProgramTakamul programSheikh Hamdan Award for Medical SciencesDana Gas Endowed Chair for Chemical EngineeringTechnology Innovation Pioneer (TIP
Continuum Modeling and Finite Element Simulation of Incompressible Dielectric Viscoelastic Actuators at Finite Strains
No full textDielectric elastomers, known for their ability to undergo large deformations exceeding 100%, are widely used as actuators in adaptive structures and soft robotics. Within the current contribution, we present a continuum material model that captures the incompressibility and viscous behavior of these polymers under finite s train and e lectric a ctuation. To address l arge deformations, we use a multiplicative decomposition of the deformation gradient to separate elastic and viscous effects. The elastic response is represented by a Yeoh potential, which is well suited to describe the material behavior under large strains. The evolution of internal strains is modeled using a dissipation function. Electric field a nd d ielectric d isplacement a re modeled i n s patial c onfiguration, le ading to an electromechanically coupled problem. We propose a mixed finite e lement f ormulation w ithin a variational framework based on the above thermodynamic principles. We introduce a novel approach using volume-preserving tensor-valued elements for internal strains, where we make use of matrix exponential functions to achieve incompressiblity exactly. As an example, we consider an experimental setup of a three-dimensional circular actuator. We provide material parameters for VHB4910 for the proposed model, and compare our results to experimental data from a different work
The Influence of Femtosecond Laser Shock Peening on the Functional Fatigue Properties of Ti₆₇Zr₁₉Nb₁₁.₅Sn₂.₅ Bio-Compatible Shape Memory Alloy
No full textA Master of Science thesis in Mechanical Engineering by Muhammad Asim entitled, “The Influence of Femtosecond Laser Shock Peening on the Functional Fatigue Properties of Ti₆₇Zr₁₉Nb₁₁.₅Sn₂.₅ Bio-Compatible Shape Memory Alloy”, submitted in November 2024. Thesis advisor is Dr. Wael Abuzaid. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).This thesis investigates the functional fatigue characteristics of Ti₆₇Zr₁₉Nb₁₁.₅Sn₂.₅, a nickel-free shape memory alloy (SMA) designed for biomedical and engineering applications, with a focus on optimizing its superelastic (SE) properties under cyclic loading conditions. Shape memory alloys, such as TiNb-based alloys, are increasingly preferred for medical applications due to their biocompatibility and superelastic properties. However, their functional fatigue performance requires further improvement for enhanced longevity in biomedical settings. This study applies femtosecond laser shock peening (F-LSP) to the chosen SMA as a novel approach to mitigate residual strain accumulation, aiming to enhance the alloy’s SE recovery and functional fatigue life. The experimental analysis involves cyclic loading tests on standard and drilled-hole specimens to replicate uniform and concentrated Stress fields. The F-LSP parameters, including laser power, scanning speed, and pulse intensity, were systematically optimized to enhance SE recovery under cyclic loading conditions. Strain mapping through Digital Image Correlation (DIC) quantified residual and recoverable strains on cyclic loading. Microstructural characterization using SEM, EDS, EBSD, and XRD revealed that F-LSP induced favourable phase transformation, which improved SE recovery without thermal degradation. The findings indicate that optimized F-LSP parameters improved the recovery of SE strains by 12% and reduced the residual strain accumulation following cyclic loading up to 25 cycles. The functional degradation of the alloy was reduced to 3.26% after F-LSP compared to the sample without LSP, which was 7.96 % after 25 loading cycles. The alloy showed apparent phase transformation at the surface from body-centred cubic (BCC) to the orthorhombic phase, potentially promoting improved superelastic recovery. These insights highlight the effectiveness of F-LSP in enhancing the functional fatigue properties of Ti-based SMAs, highlighting its potential for advancing SMA performance in high-demand biomedical applications.College of EngineeringDepartment of Mechanical EngineeringMaster of Science in Mechanical Engineering (MSME
Structural Behavior of Circular Columns Made of Ultra-High Performance Engineered Cementitious Composites (UHP-ECC)
No full textA Doctor of Philosophy Dissertation in Materials Science and Engineering by Mohamed Essam Mohamed Elkafrawy entitled, “Structural Behavior of Circular Columns Made of Ultra-High Performance Engineered Cementitious Composites (UHP-ECC)”, submitted in April 2024. Dissertation advisor is Dr. Mohammad AlHamaydeh. Soft copy is available (Dissertation, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).College of Arts and SciencesMultidisciplinary ProgramsPhD in Materials Science and Engineering (PhD-MSE
LungVision: X-ray Imagery Classification for On-Edge Diagnosis Applications
No full textThis study presents a comprehensive analysis of utilizing TensorFlow Lite on mobile phones for the on-edge medical diagnosis of lung diseases. This paper focuses on the technical deployment of various deep learning architectures to classify nine respiratory system diseases using X-ray imagery. We propose a simple deep learning architecture that experiments with six different convolutional neural networks. Various quantization techniques are employed to convert the classification models into TensorFlow Lite, including post-classification quantization with floating point 16 bit representation, integer quantization with representative data, and quantization-aware training. This results in a total of 18 models suitable for on-edge deployment for the classification of lung diseases. We then examine the generated models in terms of model size reduction, accuracy, and inference time. Our findings indicate that the quantization-aware training approach demonstrates superior optimization results, achieving an average model size reduction of 75.59%. Among many CNNs, MobileNetV2 exhibited the highest performance-to-size ratio, with an average accuracy loss of 4.1% across all models using the quantization-aware training approach. In terms of inference time, TensorFlow Lite with integer quantization emerged as the most efficient technique, with an average improvement of 1.4 s over other conversion approaches. Our best model, which used EfficientNetB2, achieved an F1-Score of approximately 98.58%, surpassing state-of-the-art performance on the X-ray lung diseases dataset in terms of accuracy, specificity, and sensitivity. The model experienced an F1 loss of around 1% using quantization-aware optimization. The study culminated in the development of a consumer-ready app, with TensorFlow Lite models tailored to mobile devices
The Impact of Poor Communication During the Pandemic Situation Within the UAE Real Estate Sector
No full textA Doctor of Philosophy Dissertation in Engineering Systems Management by Mohammed Ibrahim Mustafa Al Ustad entitled, “The Impact of Poor Communication During the Pandemic Situation Within the UAE Real Estate Sector”, submitted in April 2024. Dissertation advisor is Dr. Vian Ahmed and dissertation co-advisor is Dr. Hussam Alshraideh. Soft copy is available (Dissertation, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).Real estate is one of the most fundamental industries in the global economy. The research highlighted that poor communication is core to the project's failure. In addition, despite the literature listing numerous studies and credible scales that have been proposed over the years with information linking poor communication in construction projects, they frequently focus only on the project team and the execution stage in the third phase of the project life cycle. Moreover, there was little or no information about real estate project failure between the stakeholders. There were no defined factors of poor communication in each stage of the real estate industry's project life cycle during the pandemic research was conducted. No tool was developed to solve the issue of poor communication between the stakeholders during the project life cycle in the UAE during a pandemic, which indicates a research gap. To fill this gap, the research defines poor communication between the stakeholders throughout the project life cycle. The study focuses on finding additional factors that cause poor communication from the interviews with the stakeholders, which contribute to the finding of 58 new factors. Hence, CFA identified the most significant factor in the project life cycle, which is around 34 factors. The research developed a decision tree tool that can predict poor communication and the factors contributing to a particular stage in a real estate project. The decision support tool protects the project from any failure due to time and cost overruns supported by the risk matrix. Lastly, the study creates an application platform that will be adopted by project management in the real estate industry to test poor communication in each stage of the project life cycle in UAE during the pandemic situation. In conclusion, the research fulfills the literature review gap by developing a decision support tool that predicts the occurrence of poor communication in the real estate industry and identifies the factors that impact poor communication based on historical data in real estate during the project life cycle.College of EngineeringMultidisciplinary ProgramsPhD in Engineering - Engineering Systems Management (PhD-ESM
Cure Monitoring of Epoxy Structural Adhesion Using Lamb Waves and the Discrete Wavelet Transform
No full textEpoxy adhesives have extensive applications in the aerospace industry. In this study, the curing behaviour of preimpregnated epoxy adhesive was monitored using guided ultrasonic Lamb waves. These waves were excited and recorded using ultrasonic transducers. This study aims to analyze the recorded Lamb wave data using the discrete wavelet transform (DWT). A 5-level DWT was performed on the recorded ultrasonic signals. The approximation and detail coefficients of the signals were analyzed to extract the evolution of the degree of cure (DOC) as a function of time. The results were then validated using differential scanning calorimetry (DSC). The results have shown that the DOC obtained from the DWT matches the one obtained using DSC. This method was able to detect the start time and end time of the curing process. The correlation between the two DOCs was found to be R2 = 0.68 indicating that ultrasonic Lamb waves can be used as a real time monitoring method for the degree of cure of a material during the manufacturing process
Optimizing Energy Consumption In Cloud Datacenters
No full textA Master of Science thesis in Computer Engineering by Mueez Ahmad Khan entitled, “Optimizing Energy Consumption In Cloud Datacenters”, submitted in December 2024. Thesis advisor is Dr. Raafat Aburukba. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).Cloud computing has become a cornerstone of modern technology, enabling scalable and efficient resource utilization. However, the rapid growth in demand has resulted in significant energy consumption, posing challenges for sustainability and operational efficiency in cloud datacenters. This thesis addresses the critical issue of energy optimization in task scheduling within cloud datacenters, where increasing demand has led to significant energy consumption and environmental impact. Tasks with varying complexities are allocated to cores with unique specifications, aiming to minimize energy usage while maintaining operational efficiency. A comprehensive mathematical model is proposed to minimize energy consumption when assigning tasks to cores in a datacenter. The model is validated using exact solutions methods for small-scale instances. To further test the model on large scale problems, two hybrid heuristic algorithms based on Genetic Algorithm (HGA) and Simulated Annealing (HSA), are proposed. Parameter tuning for HGA and HSA was performed to further improve the solution quality and reduce execution time. Experiments were conducted on small, medium, large and x-large problem sets to test the scalability of the heuristics. Small size problem set was used to compare the heuristic quality to the exact solutions which showed that the heuristics provide a higher energy consumption by around 5% compared to the exact solution but with approximately 50% faster execution time. This proves that both heuristics provide a near optimal solution when compared to the exact solution with a much faster execution time. For medium-sized problems, HGA provided a lower energy consumption of around 6% over HSA, with an approximately 35% longer execution time. In large and extra-large problems, HGA outperformed HSA in providing a lower energy consumption by around 10%, but required around 23% more time for execution. This demonstrates that HSA is more suitable for scenarios where quick convergence is prioritized, whereas HGA is better suited for applications that require minimum energy consumption.College of EngineeringDepartment of Computer Science and EngineeringMaster of Science in Computer Engineering (MSCoE
Dynamic Compressive Properties of Single Crystal Multi-Principal Element Alloy (MPEA) V₁₀Fe₄₅Co₃₀Cr₁₀Ni₅
No full textA Master of Science thesis in Mechanical Engineering by Mohamed Yasser Naiem Ahmed Mohamed entitled, “Dynamic Compressive Properties of Single Crystal Multi-Principal Element Alloy (MPEA) V₁₀Fe₄₅Co₃₀Cr₁₀Ni₅”, submitted in October 2024. Thesis advisor is Dr. Wael Abuzaid and thesis co-advisor is Dr. Maen Alkhader. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).This thesis investigates the compressive behavior of the single-crystal V₁₀Fe₄₅Co₃₀Cr₁₀Ni₅ high-entropy alloy (HEA) under quasi-static and dynamic loading conditions. The alloy was tested at room temperature (RT: 298 K) for both quasi-static and dynamic loading, and at cryogenic temperature (LN: 77 K) for dynamic loading only. Four crystallographic orientations ([110], [123], [001], and [111]) were studied. The stress-strain behavior and slip system activation were characterized using Digital Image Correlation (DIC) for full-field strain mapping and Electron Backscatter Diffraction (EBSD) for microstructural analysis. Specimens were subjected to quasi-static loading at strain rates of 1.1x10⁻³ to 1.29x10⁻³ s⁻¹, while dynamic tests were conducted at strain rates of approximately 2200 to 3000 s⁻¹ using a Split Hopkinson Pressure Bar (SHPB). The results reveal significant orientation-dependent mechanical properties. Under quasi-static loading at RT, the [111] orientation exhibited the highest yield stress of 197 MPa, while the [001] orientation showed the lowest at 102 MPa. Dynamic loading at RT increased the yield strength across all orientations, with the [111] orientation reaching 305 MPa and [001] reaching 180 MPa. At cryogenic temperatures, the yield strength further increased, with the [111] orientation achieving 448 MPa. Despite the significant increase in strength at cryogenic deformation temperatures, the considered material still exhibited a notable ductile response. Slip-dominated deformation was observed in all orientations. No twinning-induced plasticity (TWIP) or transformation-induced plasticity (TRIP) were observed, indicating that slip was the primary deformation mechanism under all conditions. These findings provide valuable insights into the performance of the V₁₀Fe₄₅Co₃₀Cr₁₀Ni₅ HEA, particularly under high strain-rate and cryogenic conditions.College of EngineeringDepartment of Mechanical EngineeringMaster of Science in Mechanical Engineering (MSME