Foundation University Journal of Engineering and Applied Sciences
Not a member yet
56 research outputs found
Sort by
Pictorial Task Assistance System using Electroencephalography Signals
Get PDFNeuromuscular disorders are a significant health problem globally. Patients may experience paralysis, muscle weakness, and communication problems because of these disorders. We propose a Pictorial Task Assistance System to help patients with communication issues using Electroencephalography (EEG). We developed an interface for patients containing an image of food and water. We collected EEG data from 25 healthy students using the Muse headset and Muse monitor app for our study, while they selected one of the images. The EEG data was used to train three supervised machine-learning algorithms for classification. The labels were acquired manually from participants. Using 10-fold cross-validation the results demonstrated that the Random Forest (RF) classifier achieved 88% accuracy, K-Nearest Neighbors (KNN) 80%, and 76% accuracy in Logistic Regression (LR) in the classification of food and water images. These results suggest that the proposed system has the potential to be a useful tool for patients suffering from neuromuscular disorders to perform communication for their necessary tasks
Dynamic and Integrated Security Model in Inter Cloud for Image Classification
Get PDFCloud computing has transformed software and database accessibility, utilizing the Internet and server hosting. However, security risks arise, including malware attacks and website hacking. To address these challenges, deep learning models like ResNet50 have been developed. Trained on encrypted images, ResNet50 enhances the speed and accuracy of image recognition, enabling the identification of hidden data without decryption. Despite inter-cloud communication issues, cloud servers prioritize data security, user privacy, and integrity maintenance. The ResNet50 model exhibits impressive performance, achieving 99.5%accuracy and precision-recall scores of 99.5% and 99.5% using the ImageNet Dataset. Cloud computing offers significant advantages, but data security remains a critical concern. Encrypted image recognition powered by deep learning models offers efficient and private solutions. Cloud providers continually strive to improve inter-cloud communication, ensuring comprehensive protection for data and system integrity. The remarkable capabilities of ResNet50 highlight its potential in encrypted image analysis tasks
Securing the Internet of Things: A Comprehensive Review of Security Challenges and Artificial Intelligence Solutions
Get PDFOne of the major needs and challenges of this century is the use of cutting-edge technology considering the industry 4.0 revolution. The Internet of Things (IoT) falls in the category of a cutting-edge example of such innovation in the computing and information industry. In IoT compared to classical networking methods practically; every device we employ is accessible at any time from any location. Nevertheless, IoT continues to encounter several security challenges, and the magnitude of cyber-physical security risks is escalating alongside the widespread use of IoT technologies considering Moore’s laws expected to be 30 billion devices by 2025. IoT will continue to face vulnerabilities and risks unless there is a comprehensive understanding and proactive approach towards tackling its security concerns. To ensure both the cyber and physical security of IoT devices during data gathering and sharing, it is imperative to evaluate security considerations, identify instances of cyber-attacks, and implement effective security protocols at multiple layers for making highly secured IoT. Conventional security measures like data classification, strict access controls, monitoring privileged account access, encrypting sensitive data, security awareness training, network segregation, segmentation cloud security, application security, patch management, and physical security employed in the realm of IoT are inadequate in light of the current security difficulties posed by the proliferation of sophisticated attacks and threats. Utilization of artificial intelligence (AI) techniques, especially machine and deep learning models is becoming a compelling and effective approach to enhance security of the IoT devices. This research article presents a comprehensive review of the key aspects of IoT security, including the challenges, potential opportunities, and AI-driven solutions. The primary goal of this article is to provide technical resources for cybersecurity experts and researchers working on IoT initiatives
Exploring the Art of Sampling and Reconstruction - A PBL Approach to Signal Processing
Get PDFDigital signal processing has facilitated the digital representation, analysis, and transmission of analog signals. This work presents a Project-Based Learning (PBL) approach to encourage students to work on real-world projects or challenges to gain knowledge and skills in the field of signal sampling and reconstruction, focusing on their significance in multidimensional domains where they are applied, such as communication systems, image processing, or audio signal processing. Sampling is how a continuous-time signal is transformed into a discrete format, i.e., when we select values at different time points. This requires taking samples of signal amplitudes at uniformly spaced intervals, which creates a stream of quantities. But the main difference is that to extract information from an analog signal, we need samples; there is no other way. This process is referred to as sampling rate frequency, so it is the number of samples collected during some period of time. Reconstruction, on the other hand, means doing a conversion of it from time-discrete to its continuous-time form. This operation generated an approximated signal that is continuous from those sampled values. Different types of reconstruction methods, such as ideal interpolation, zero-order hold, or since interpolation, are chosen based on signal features and need. Given that the reconstruction process is limited because it is based on only a finite window of samples, it becomes clear how important accurate sampling and reconstruction are in preserving the original quality of a signal, minimizing distortion during this part of the audio chain. The Nyquist–Shannon sampling theorem, also called as Nyquist criterion or sometimes as Shannon sampling theorem, defines a good minimum rate at which a band-limited signal to be sampled so that it can be reconstructed without the loss of information. It is important to note that it would have a big effect on the systems that can be developed, and that are both efficient and dependable. In brief, learning sampling and reconstruction is arguably the most basic of signal processing concepts. Appropriate sampling allows to keep the integrity and quality of a signal across a wide range of applications, i.e., from new communication technologies with diminished bandwidth, multimedia systems to many other subjects that aim for innovative high-performing digital systems
MEREC-WISP(S) Integration Extended with Fermatean Fuzzy Set for Requirement Prioritization
Get PDFSoftware projects need efficient requirement prioritization. Time, budget, and quality often limit these projects. MCDM techniques help balance conflicting criteria. However, they struggle to rank options due to multiple parties. Current MCDM methods have drawbacks, like poor uncertainty management. This thesis presents a new technique, WISP-S, for requirement prioritization in software development. The dynamic WISP(S) approach and the MEREC method are merged. Fermatean fuzzy numbers manage qualitative data and uncertainty. This technique surpasses the restrictions of current MCDM methods. It offers an effective way to rank software requirements within limits. The research consists of two separate stages. The first phase introduces an enhanced MEREC technique. It’s designed to compute the objective weights of each criteria within Fermatean Fuzzy Sets (FFS). By leveraging FFS features, this expansion improves the current MEREC technique. It allows a more thorough examination of criteria weights. The second phase integrates the WISP(S) method with the suggested generalized weighted Fermatean fuzzy aggregated operator and MEREC technology. This integration allows ranking and evaluating alternatives in a prioritization context. The study offers a robust strategy to prioritize alternatives. It considers both qualitative and uncertain data by integrating two novel methods. The approach’s validity and robustness are confirmed through comparisons with existing models and sensitivity analysis. A real-world case study demonstrates the proposed approach. The WISP-S approach is a novel technique for integrated Fermatean fuzzy information-based decision-making. It can handle complex decision-making problems in software development. This approach could improve the success of software projects. It addresses the shortcomings of existing MCDM methods
Software Level Security, Privacy Attacks and Challenges in Smart Healthcare Systems
Get PDFThe widespread adoption of smart healthcare systems and the use of IoT, cloud computing, robotics, and Artificial Intelligence for pandemic data analysis, remote diagnostics, the use of wireless medical devices, and public information services has introduced new cybersecurity risks. The vast amount of personal data, including health information, makes healthcare systems a prime target for cybercriminals who could exploit it for harassment and fraudulent purposes. Additionally, the increased reliance on anytime, anywhere connectivity creates vulnerabilities in wireless medical devices. Recent reports highlight these threats, with examples including privacy breaches, ransomware attacks, and disruptions in communication channels for medical devices. Such security breaches can erode patient trust, cripple healthcare systems, and even endanger lives. This surge in usage and integration of smart devices, often connected through healthcare apps, has also introduced new cybersecurity concerns. To address these challenges, this paper reviews the recent software-level security and privacy threats in smart healthcare systems, explores mitigation techniques proposed in the literature, and discusses the vulnerabilities of medical devices along with the service disruption in the communication channel, and the impact of cybersecurity attacks on smart healthcare systems, and discusses the future challenges of software security in smart healthcare
Artificial Intelligence (AI) Driven Automated Learning Management System
Get PDFLearning management platforms are widely used throughout educational institutes worldwide and have become an industry standard. These systems provide an easy way for students and teachers to communicate and disseminate information, but this can be a cumbersome task that takes up a lot of time and effort at the teacher’s end. With the evolution in artificial intelligence, this system can be made intelligent, which could really help teachers and educators better manage their workload and maintain a better work-life balance while focusing more on students rather than creating quizzes and assignments. We are proposing a framework that solves all these issues by generating content required by educators and students, such as assignments, quizzes, and lecture notes, automatically, without putting in minimal to no manual effort. The proposed framework will undergo rigorous testing so its performance, usability, efficiency, and safety can be measured to make sure that the system is effective yet safe to use in a real-world environment
Strategic Analysis of Feature Selection Methods for Enhanced Dental Therapy Recognition in Machine Learning Applications
Get PDFThe popularity of smart clinics has increased significantly as a result of technical developments in fields like computer vision. At the heart of such systems is the ability to recognize objects and activities as well as perceive the environment as a whole. This is crucial for both eco-independent systems and human-machine interaction, especially in settings with constrained workspaces, like dental care. Our study delves into an extensive analysis of multiple machine learning models designed to robustly predict dental treatments. These models encompass Lazy Instance-based Learning, Sequential Minimal Optimization, Hoeffding Tree, and Random Tree. Leveraging object-oriented input sourced from gaze-guided wearable cameras, we scrutinize intricate attributes such as material properties, patients' dental conditions, and the array of instruments in use. Notably, we exploit the insight that identifying visual cues during an activity holds the potential to address the specific therapy identification challenge. Utilizing a dental data set that we gathered in the real world, we conducted our experiments and discovered that combining multiple criteria enhances accuracy in comparison to using each one alone. We did see, nevertheless, that in certain circumstances employing the symptoms alone produced superior outcomes. Additionally, symptoms demonstrated to have lesser error than combination in terms of RMS error convergence. Finally, we observed that the machine learning models' build and test durations increased as a result of the combined method. This demonstrates that adding additional parameters does not necessarily result in better outcomes in machine learning applications generally and in medical/dental applications in particular. Instead, it relies on the machine learning tool used, the settings taken into account and the input data provided. The versatility of our approach extends beyond dental contexts. It has been systematically validated across diverse domains, including the recognition of kitchen activities within smart home environments. This methodology holds relevance for various outdoor scenarios where the focal point of attention guides ongoing activities
Towards Mining Variable Features in Software Product Lines during Development
Get PDFSoftware Product Line (SPL) engineering provides a strong vision to develop highly adaptable software systems using the common characteristics as well as controlling the variabilities of a product line of products. The specification of some variable features and their exact reuse is, however, a major problem, in particular when handling heterogeneous families of products and concurring multi-facet demands of the stakeholders. It is complicated by the fact that approaches to managing commonality and variability are not systematic in the early phases of requirement specification. The current feature selection algorithms tend not to be rigorous enough to deal with multi-stakeholder viewpoints and fully categorize features by their natural facets, which makes them inefficient and impairs their reusability. To constrain this serious issue of research, a new, articulate model of variable feature mining and selection in SPLs is given in this paper. The distinctive points of our methodology are that we combine systematic requirements collecting, rigorous data preprocessing, and an unusual aspect-based feature clustering strategy based on using unsupervised learning algorithms (We are using Weka, for instance). This feature-conscious classification with subsequent learning through supervision to identify two types of features, Common and Variable ones, distinguishes our model over the traditional, simpler divide-and-conquer methods due to a more accurate and context-situated feature taxonomy. Two industrial studies of a biometric system and an online auction system were conducted in a rigorous manner to assess the proposed model. Early findings have shown that the model suggested is of critical importance in improving the procedure of variable feature selection. Major discoveries showed significant advances in the effective accuracy of classification features, significant advancements in the efficiency of the feature selection (e.g., less time and effort than manual processes involved), and raised the satisfaction levels of the stakeholders with the chosen feature sets. The study helps practitioners in the industry because it provides them with a data-driven, structured approach that enhances the feature selection task, in addition to the fact that the study helps deal with reusability and customization across the SPL development life-cycle successfully. Finally, this model, combined with the earlier elements, helps to arrive at a solution that is time-to-market quicker and also yields product configurations that are stronger and meet a fundamental need of existing SPL practices
Discerning of Hall Effect in Technology using New Data Acquisition Technique
Get PDFCertain areas of physics are revolutionized by modern technology, and the Hall Effect Apparatus (HEA) is one of those, as this apparatus has a lot of importance. To study the HE apparatus, an electronics-based system is designed through which one can be informed about the Hall Voltage (VH), Hall Coefficient (RH), and the type of majority charge carriers in semiconductor samples using microcontrollers with a built-in graphic user interface (GUI) as a stand-alone system. The purpose is to develop and upgrade the HE apparatus using advanced data acquisition techniques and introduce microcontrollers for a GUI to obtain electronic data and graphical representations. The measured results show that the system can produce 90% results at 10 times lower cost. Generally, three semiconductor samples were taken to observe the Hall Effect parameters using the HE Apparatus, and remarkable relative results with the literature were observed. Carrier concentration for CdHgTe, ZnO, and Silicon were -10×1020/cm3, -2×1018/cm3, and 7.5×1018/cm3, respectively, using the I/V slope from the plots obtained at different magnetic fields. Similarly, the Hall coefficients, Hall Mobility, and the Resistivity