arXiv subjects

M. A. Sayed

Publications and source records attributed to M. A. Sayed.

3 recordsLinked to original sources

arXiv2023

Prediction of Pneumonia and COVID-19 Using Deep Neural Networks

Pneumonia, caused by bacteria and viruses, is a rapidly spreading viral infection with global implications. Prompt identification of infected individuals is crucial for containing its transmission. This study explores the potential of medical image analysis to address this challenge. We propose machine-learning techniques for predicting Pneumonia from chest X-ray images. Chest X-ray imaging is vital for Pneumonia diagnosis due to its accessibility and cost-effectiveness. However, interpreting X-rays for Pneumonia detection can be complex, as radiographic features can overlap with other respiratory conditions. We evaluate the performance of different machine learning models, including DenseNet121, Inception Resnet-v2, Inception Resnet-v3, Resnet50, and Xception, using chest X-ray images of pneumonia patients. Performance measures and confusion matrices are employed to assess and compare the models. The findings reveal that DenseNet121 outperforms other models, achieving an accuracy rate of 99.58%. This study underscores the significance of machine learning in the accurate detection of Pneumonia, leveraging chest X-ray images. Our study offers insights into the potential of technology to mitigate the spread of pneumonia through precise diagnostics.

eess.IV↗

arXiv2022

Investigating the Security of EV Charging Mobile Applications As an Attack Surface

In this paper, we study the security posture of the EV charging ecosystem against a new type of remote that exploits vulnerabilities in the EV charging mobile applications as an attack surface. We leverage a combination of static and dynamic analysis techniques to analyze the security of widely used EV charging mobile applications. Our analysis was performed on 31 of the most widely used mobile applications including their interactions with various components such as cloud management systems. The attack, scenarios that exploit these vulnerabilities were verified on a real-time co-simulation test bed. Our discoveries indicate the lack of user/vehicle verification and improper authorization for critical functions, which allow adversaries to remotely hijack charging sessions and launch attacks against the connected critical infrastructure. The attacks were demonstrated using the EVCS mobile applications showing the feasibility and the applicability of our attacks. Indeed, we discuss specific remote attack scenarios and their impact on EV users. More importantly, our analysis results demonstrate the feasibility of leveraging existing vulnerabilities across various EV charging mobile applications to perform wide-scale coordinated remote charging/discharging attacks against the connected critical infrastructure (e.g., power grid), with significant economical and operational implications. Finally, we propose countermeasures to secure the infrastructure and impede adversaries from performing reconnaissance and launching remote attacks using compromised accounts.

cs.CR↗

arXiv2021

On sub-grid scale modeling in a differentially heated cavity using coarse large eddy simulation

Two widely used sub-grid scale models, the standard and the dynamic Smagorinsky models were tested in a simulation of the flow in a thermally driven 3D cavity at Rayleigh number of Ra=1E9. The main focus of the research is the response of the subgrid scale models to the coarse mesh and their ability to predict the flow in a thermally driven 3D cavity with the mesh resolution below generally accepted LES standards. The research is motivated by investigating a feasible modeling strategy for the particulate flow in the wall-bounded side-wall heated cavity. As URANS and hybrid RANS-LES models fail to produce modeled energy due to laminarization, the alternative is LES applied on the coarse mesh. In a quantitative manner, first and second-moment statistics are com-pared against both reference LES and experimental databases from literature. For first moment statistics, both models globally predict the flow field well. However, for higher moments, the dynamic model outperforms the standard Smagorinsky model using the same mesh resolution. In a qualitative fashion, we report the contours of both temperature and velocity fields as well as the near-wall turbulent coherent structures. Results of flow quantities are globally in very good agreement with both reference LES and experimental measurements at a fraction of the CPU power needed for conventional LES.

physics.flu-dyn↗