Mudassar Ali Khan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mudassar Ali Khan has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Ikram Ud Din
Department of Information Technology, The University of Haripur, Haripur, Pakistan
13 shared publications - Ahmad Almogren
Department of Computer Science, College of Computer and Information Sciences, King Saud University
12 shared publications - Nauman Ali Khan
Department of Computer Software Engineering, National University of Sciences and Technology, Islamabad, Pakistan
5 shared publications - Byung-Seo Kim
Department of Software and Communications Engineering, Hongik University, Sejong 30016, Republic of Korea
3 shared publications - Fahim Arif
Department of Computer Software Engineering, National University of Sciences and Technology, Islamabad, Pakistan
2 shared publications - Joel Jose Puga Coelho Rodrigues
Federal University of Piau (UFPI), Teresina - PI, Brazil
2 shared publications
Latest Publications
- 23 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (2) , 3677-3684
Diabetic Retinopathy Classification for Consumer Electronics Using DenseNet and Grad-CAM++
Diabetic retinopathy (DR) is a progressive eye disorder that may lead to permanent blindness when it is not treated. Nevertheless, the issue of early recognition is a problem since the retinal assessment takes a lot of time, and ophthalmologists are in short supply in most places. The paper fulfills this urgent requirement by suggesting a light-weight deep learning architecture to classify DR, namely, in the consumer electronics domain. The model combines DenseNet models and Grad-CAM++ to enhance both interpretation and extraction of features. It involves three dense net models that are used to label retinal images, D121- [M1], D169- [M2], and D121- [M3] to classify APTOS-2019 retinal image samples, 44570 sample images of 5 levels of DR. The framework produces detailed heatmaps which indicate significant parts of the retina which affect classification through incorporation of Grad-CAM++. The design is made to comply with the requirements of consumer devices with a focus on efficiency and flexibility. The steps involved in preprocessing are image resizing to 224x224 pixels and data augmentation to enhance strength. D121-[M3] was the best performing of the models; it received 95.63 percent training accuracy, 0.9921 precision, 0.9930 recall, and a 0.9925 F1-score.
- 23 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (2) , 3677-3684
Diabetic Retinopathy Classification for Consumer Electronics Using DenseNet and Grad-CAM++
Diabetic retinopathy (DR) is a progressive eye disorder that may lead to permanent blindness when it is not treated. Nevertheless, the issue of early recognition is a problem since the retinal assessment takes a lot of time, and ophthalmologists are in short supply in most places. The paper fulfills this urgent requirement by suggesting a light-weight deep learning architecture to classify DR, namely, in the consumer electronics domain. The model combines DenseNet models and Grad-CAM++ to enhance both interpretation and extraction of features. It involves three dense net models that are used to label retinal images, D121- [M1], D169- [M2], and D121- [M3] to classify APTOS-2019 retinal image samples, 44570 sample images of 5 levels of DR. The framework produces detailed heatmaps which indicate significant parts of the retina which affect classification through incorporation of Grad-CAM++. The design is made to comply with the requirements of consumer devices with a focus on efficiency and flexibility. The steps involved in preprocessing are image resizing to 224x224 pixels and data augmentation to enhance strength. D121-[M3] was the best performing of the models; it received 95.63 percent training accuracy, 0.9921 precision, 0.9930 recall, and a 0.9925 F1-score.
- 31 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71 (4) , 11761-11768
TALB—Trust-Aware Load Balancing Technique for Improved Communication in Consumer IoT
The ultra-dense node scaling of Internet of Things (IoT) deployments necessitates trust-aware and load-resilient routing mechanisms for secure and efficient network coordination within distributed, resource-constrained environments. Existing frameworks inadequately integrate dynamic trust metrics with load-balancing decisions, often resulting in node overload and unstable task delegation. To address this, this work presents a trust-driven, load-balanced routing scheme aimed at securing task assignment under distributed IoT conditions, including consumer-centric IoT systems. The core framework comprises four integrated modules: (i) a Communication Trust Computation Engine that combines entropy-weighted integrity and reliability scores to derive adaptive trust estimates; (ii) a Trust-Load Synchronization Protocol employing noise-resilient status exchange and fault-triggered resynchronization mechanisms; (iii) an Adaptive Routing and Task Assignment Engine that applies composite eligibility masks, entropy-penalized capacity estimators, and fallback activation; and (iv) a Trust Score Recovery and Node Reintegration Module using entropy-guided revalidation cycles to reinstate penalized nodes. Simulation on the TON_IoT dataset under 15 uncertainty scenarios shows TALB (Trust-Aware Load Balancing) achieves 94.2% Uncertainty Handling Efficiency, 96.1% accuracy, and maintains latency under 155 ms with scalability validated across 1250-node deployments for home-automation and consumer IoT networks.
- 31 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71 (4) , 11761-11768
TALB—Trust-Aware Load Balancing Technique for Improved Communication in Consumer IoT
The ultra-dense node scaling of Internet of Things (IoT) deployments necessitates trust-aware and load-resilient routing mechanisms for secure and efficient network coordination within distributed, resource-constrained environments. Existing frameworks inadequately integrate dynamic trust metrics with load-balancing decisions, often resulting in node overload and unstable task delegation. To address this, this work presents a trust-driven, load-balanced routing scheme aimed at securing task assignment under distributed IoT conditions, including consumer-centric IoT systems. The core framework comprises four integrated modules: (i) a Communication Trust Computation Engine that combines entropy-weighted integrity and reliability scores to derive adaptive trust estimates; (ii) a Trust-Load Synchronization Protocol employing noise-resilient status exchange and fault-triggered resynchronization mechanisms; (iii) an Adaptive Routing and Task Assignment Engine that applies composite eligibility masks, entropy-penalized capacity estimators, and fallback activation; and (iv) a Trust Score Recovery and Node Reintegration Module using entropy-guided revalidation cycles to reinstate penalized nodes. Simulation on the TON_IoT dataset under 15 uncertainty scenarios shows TALB (Trust-Aware Load Balancing) achieves 94.2% Uncertainty Handling Efficiency, 96.1% accuracy, and maintains latency under 155 ms with scalability validated across 1250-node deployments for home-automation and consumer IoT networks.
- 23 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. PP (21682194) , 1-8
Interpretable and Adaptive GAN-BiLSTM Approach for Cyber Threat Detection in IoMT-based Healthcare 5.0
Healthcare 5.0, driven by the Internet of Medical Things (IoMT), introduces transformative changes in the medical field but also exposes systems to growing cybersecurity threats. While Deep Learning (DL) offers high accuracy in attack detection, its effectiveness is often limited by data imbalance and difficulty in identifying key features dynamically. Additionally, DL models are often criticized for their lack of interpretability, as their internal decisionmaking remains obscure. To overcome these limitations, this paper presents an explainable and adaptive DL-based security framework. It integrates a Generative Adversarial Network (GAN) to balance the dataset by generating realistic samples for underrepresented attack classes, and employs Bidirectional Long Short-Term Memory (BiLSTM) to identify temporal patterns and critical features. To enhance transparency, SHapley Additive exPlanations (SHAP) and Permutation Feature Importance (PFI) are used for interpreting the model's decisions. Experiments conducted on the NSL-KDD dataset demonstrate the effectiveness of the proposed method, achieving 93.81% accuracy and an F1-score of 82.95%.
- 23 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. PP (21682194) , 1-8
Interpretable and Adaptive GAN-BiLSTM Approach for Cyber Threat Detection in IoMT-based Healthcare 5.0
Healthcare 5.0, driven by the Internet of Medical Things (IoMT), introduces transformative changes in the medical field but also exposes systems to growing cybersecurity threats. While Deep Learning (DL) offers high accuracy in attack detection, its effectiveness is often limited by data imbalance and difficulty in identifying key features dynamically. Additionally, DL models are often criticized for their lack of interpretability, as their internal decisionmaking remains obscure. To overcome these limitations, this paper presents an explainable and adaptive DL-based security framework. It integrates a Generative Adversarial Network (GAN) to balance the dataset by generating realistic samples for underrepresented attack classes, and employs Bidirectional Long Short-Term Memory (BiLSTM) to identify temporal patterns and critical features. To enhance transparency, SHapley Additive exPlanations (SHAP) and Permutation Feature Importance (PFI) are used for interpreting the model's decisions. Experiments conducted on the NSL-KDD dataset demonstrate the effectiveness of the proposed method, achieving 93.81% accuracy and an F1-score of 82.95%.
- 13 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (22) , 46075-46082
Hybrid CNN-LSTM Model for DDoS Attack Detection in Internet of Things-Based Healthcare Industry 5.0
The convergence of the Internet of Things (IoT) and software-defined networking (SDN) has paved the way for a new technological paradigm in Healthcare Industry 5.0. This integration addresses the complexity, heterogeneity, and dynamic nature of smart IoT devices within healthcare systems. However, it also increases the risk of cyberattacks, particularly Distributed Denial of Service (DDoS) attacks, which pose significant threats to such critical infrastructure. While deep learning (DL)-based intrusion detection methods have demonstrated high accuracy in detecting these attacks, their opaque decision-making process often leads to their characterization as black-box models, limiting their practical use for security analysts. To overcome these challenges, this study proposes an explainable hybrid model for DDoS attack detection in SDN-IoT-based Healthcare Industry 5.0 environments. The model combines the strengths of convolutional neural networks (CNNs) for spatial feature extraction and long short-term memory (LSTM) networks for capturing temporal dependencies in network traffic. Implemented using an SDN controller, the model accurately classifies DDoS and IoT attacks while providing transparency through the shapley additive explanation (SHAP) method, which identifies the most influential features in the model’s decision-making process. Simulation results on the CICDDoS2019 and IoT Healthcare Security datasets demonstrate the model’s effectiveness, achieving detection accuracy of 99.59% and 98.12%, respectively. These findings confirm the robustness of the proposed hybrid model compared to state-of-the-art methods for detecting potential attacks in Healthcare Industry 5.0 systems.
- 13 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (22) , 46075-46082
Hybrid CNN-LSTM Model for DDoS Attack Detection in Internet of Things-Based Healthcare Industry 5.0
The convergence of the Internet of Things (IoT) and software-defined networking (SDN) has paved the way for a new technological paradigm in Healthcare Industry 5.0. This integration addresses the complexity, heterogeneity, and dynamic nature of smart IoT devices within healthcare systems. However, it also increases the risk of cyberattacks, particularly Distributed Denial of Service (DDoS) attacks, which pose significant threats to such critical infrastructure. While deep learning (DL)-based intrusion detection methods have demonstrated high accuracy in detecting these attacks, their opaque decision-making process often leads to their characterization as black-box models, limiting their practical use for security analysts. To overcome these challenges, this study proposes an explainable hybrid model for DDoS attack detection in SDN-IoT-based Healthcare Industry 5.0 environments. The model combines the strengths of convolutional neural networks (CNNs) for spatial feature extraction and long short-term memory (LSTM) networks for capturing temporal dependencies in network traffic. Implemented using an SDN controller, the model accurately classifies DDoS and IoT attacks while providing transparency through the shapley additive explanation (SHAP) method, which identifies the most influential features in the model’s decision-making process. Simulation results on the CICDDoS2019 and IoT Healthcare Security datasets demonstrate the model’s effectiveness, achieving detection accuracy of 99.59% and 98.12%, respectively. These findings confirm the robustness of the proposed hybrid model compared to state-of-the-art methods for detecting potential attacks in Healthcare Industry 5.0 systems.
- Open Access
- 7 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 149272-149287
Enhancing Hyperspectral Image Analysis With Deep Learning: An Innovative Hybrid Approach for Improved Detection and Classification Using Deep-Detect
Surface material identification using hyperspectral imaging (HSI) analysis is a crucial and challenging issue in remote sensing. Researchers widely recognize that exploiting spectral-spatial data outperforms using spectral pixel-wise methods. Most of the work focuses on machine learning techniques to address issues with HSI categorization. The semi-supervised learning model, for example, was utilized in early research to address the issue of the expenses associated with training samples during HSI classification. Nevertheless, these approaches and techniques have certain drawbacks. The suggested framework must be improved in terms of hardware and computational capacity to enhance security. While integrated patch-based convolutional neural network (CNN) can help overcome these issues, due to the complicated geographical distribution of many targets, there are some downsides, such as using spectra as odorless vectors. The supervised learning techniques categorize the inputs by using a collection of appropriate examples for each class, referred to as training samples. For this purpose, samples must be obtained, however, the process is expensive and time-consuming. Deep learning techniques for evaluating spatial spectrum data have significantly advanced, yet security challenges remain in remote sensing with HSI. In this article, we introduce a novel HSI classification method, termed Deep-Detect, which leverages deep learning models with attention mechanisms to tackle these challenges. We use two weighting factors that may be obtained flexibly during the training phase to integrate both spectral and spatial properties, whose importance differs for various objects and scenarios. Additionally, we proposed a spatial and spectral fully convolutional networks (Spe-FCNs) attention mechanism that focuses attention on the input patches’ most informative regions. To verify performance, we used the northern Italy dataset from Pavia University and compared our proposed model with state-of-the-art HSI classification models. Our findings show that the proposed Deep-Detect model enhances classification accuracy while addressing key security concerns in HSIs.
- 7 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 149272-149287
Open AccessEnhancing Hyperspectral Image Analysis With Deep Learning: An Innovative Hybrid Approach for Improved Detection and Classification Using Deep-Detect
Surface material identification using hyperspectral imaging (HSI) analysis is a crucial and challenging issue in remote sensing. Researchers widely recognize that exploiting spectral-spatial data outperforms using spectral pixel-wise methods. Most of the work focuses on machine learning techniques to address issues with HSI categorization. The semi-supervised learning model, for example, was utilized in early research to address the issue of the expenses associated with training samples during HSI classification. Nevertheless, these approaches and techniques have certain drawbacks. The suggested framework must be improved in terms of hardware and computational capacity to enhance security. While integrated patch-based convolutional neural network (CNN) can help overcome these issues, due to the complicated geographical distribution of many targets, there are some downsides, such as using spectra as odorless vectors. The supervised learning techniques categorize the inputs by using a collection of appropriate examples for each class, referred to as training samples. For this purpose, samples must be obtained, however, the process is expensive and time-consuming. Deep learning techniques for evaluating spatial spectrum data have significantly advanced, yet security challenges remain in remote sensing with HSI. In this article, we introduce a novel HSI classification method, termed Deep-Detect, which leverages deep learning models with attention mechanisms to tackle these challenges. We use two weighting factors that may be obtained flexibly during the training phase to integrate both spectral and spatial properties, whose importance differs for various objects and scenarios. Additionally, we proposed a spatial and spectral fully convolutional networks (Spe-FCNs) attention mechanism that focuses attention on the input patches’ most informative regions. To verify performance, we used the northern Italy dataset from Pavia University and compared our proposed model with state-of-the-art HSI classification models. Our findings show that the proposed Deep-Detect model enhances classification accuracy while addressing key security concerns in HSIs.
- Open Access
- 16 May 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (10) , 8120
Visualization of Remote Patient Monitoring System Based on Internet of Medical Things
Remote patient monitoring (RPM) has become a crucial tool for healthcare professionals in the monitoring and management of patients, particularly for patients with chronic illnesses. RPM has undergone improvements in its capability to deliver real-time data and information to healthcare practitioners as the Internet of Medical Things (IoMT) devices have become more widely available. However, managing and analyzing such a large volume of data still remains a difficult task. The visualization method suggested in this article enables healthcare professionals to examine data gathered by IoMT devices in real-time. Healthcare professionals may monitor patient health status and identify any data irregularities thanks to the system’s dashboard. To assess the system’s usability and user satisfaction, we employed both the Post-Study System Usability Questionnaire (PSSUQ) and the System Usability Scale (SUS). The outcomes of the PSSUQ and SUS assessments revealed that the suggested visualization system scored higher than the control group, demonstrating the system’s usability, accuracy, and dependability as well as its user-friendliness and intuitive interface. The visualization system can boost the effectiveness and efficiency of remote patient monitoring, resulting in better patient care and lower healthcare costs.
- 16 May 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (10) , 8120
Open AccessVisualization of Remote Patient Monitoring System Based on Internet of Medical Things
Remote patient monitoring (RPM) has become a crucial tool for healthcare professionals in the monitoring and management of patients, particularly for patients with chronic illnesses. RPM has undergone improvements in its capability to deliver real-time data and information to healthcare practitioners as the Internet of Medical Things (IoMT) devices have become more widely available. However, managing and analyzing such a large volume of data still remains a difficult task. The visualization method suggested in this article enables healthcare professionals to examine data gathered by IoMT devices in real-time. Healthcare professionals may monitor patient health status and identify any data irregularities thanks to the system’s dashboard. To assess the system’s usability and user satisfaction, we employed both the Post-Study System Usability Questionnaire (PSSUQ) and the System Usability Scale (SUS). The outcomes of the PSSUQ and SUS assessments revealed that the suggested visualization system scored higher than the control group, demonstrating the system’s usability, accuracy, and dependability as well as its user-friendliness and intuitive interface. The visualization system can boost the effectiveness and efficiency of remote patient monitoring, resulting in better patient care and lower healthcare costs.
- Open Access
- 15 March 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (6) , 5207
Securing Access to Internet of Medical Things Using a Graphical-Password-Based User Authentication Scheme
Digital healthcare services have seen significant growth in this decade and many new technologies have been thoroughly examined to provide efficient services through secure infrastructures. The Internet of Medical Things (IoMT) revitalizes a healthcare infrastructure by creating an interconnected, intelligent, accessible, and efficient network. While there have been many studies on possible device authentication techniques for the IoMT, there is still much work to be done in user authentication to provide sustainable IoT solutions. Graphical passwords, which use visual content such as images instead of traditional text-based passwords, can help users authenticate themselves. However, current schemes have limitations. Therefore, this paper proposes a novel graphical authentication scheme that uses multiple factors to register and authenticate users using simple arithmetic operations, machine learning for hand gesture recognition, and medical images for recall purposes. The proposed method is designed to keep the authentication process simple, memorable, and robust. To evaluate the proposed scheme, we use the Post-Study System Usability Questionnaire (PSSUQ) to compare it with PIN-based and pattern-based authentication techniques. While comparing treatment and comparison groups, system quality showed a 16.7% better score, information quality a 25% increase, interface quality a 40% increase, and overall quality showed a 25% increase. The proposed method successfully revitalizes the use of graphical passwords, specifically in the field of IoMT, by developing a user-friendly, satisfying, and robust authentication scheme.
- 15 March 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (6) , 5207
Open AccessSecuring Access to Internet of Medical Things Using a Graphical-Password-Based User Authentication Scheme
Digital healthcare services have seen significant growth in this decade and many new technologies have been thoroughly examined to provide efficient services through secure infrastructures. The Internet of Medical Things (IoMT) revitalizes a healthcare infrastructure by creating an interconnected, intelligent, accessible, and efficient network. While there have been many studies on possible device authentication techniques for the IoMT, there is still much work to be done in user authentication to provide sustainable IoT solutions. Graphical passwords, which use visual content such as images instead of traditional text-based passwords, can help users authenticate themselves. However, current schemes have limitations. Therefore, this paper proposes a novel graphical authentication scheme that uses multiple factors to register and authenticate users using simple arithmetic operations, machine learning for hand gesture recognition, and medical images for recall purposes. The proposed method is designed to keep the authentication process simple, memorable, and robust. To evaluate the proposed scheme, we use the Post-Study System Usability Questionnaire (PSSUQ) to compare it with PIN-based and pattern-based authentication techniques. While comparing treatment and comparison groups, system quality showed a 16.7% better score, information quality a 25% increase, interface quality a 40% increase, and overall quality showed a 25% increase. The proposed method successfully revitalizes the use of graphical passwords, specifically in the field of IoMT, by developing a user-friendly, satisfying, and robust authentication scheme.
- Open Access
A Survey of Authentication in Internet of Things-Enabled Healthcare Systems
The Internet of medical things (IoMT) provides an ecosystem in which to connect humans, devices, sensors, and systems and improve healthcare services through modern technologies. The IoMT has been around for quite some time, and many architectures/systems have been proposed to exploit its true potential. Healthcare through the Internet of things (IoT) is envisioned to be efficient, accessible, and secure in all possible ways. Even though the personalized health service through IoT is not limited to time or location, many associated challenges have emerged at an exponential pace. With the rapid shift toward IoT-enabled healthcare systems, there is an extensive need to examine possible threats and propose countermeasures. Authentication is one of the key processes in a system’s security, where an individual, device, or another system is validated for its identity. This survey explores authentication techniques proposed for IoT-enabled healthcare systems. The exploration of the literature is categorized with respect to the technology deployment region, as in cloud, fog, and edge. A taxonomy of attacks, comprehensive analysis, and comparison of existing authentication techniques opens up possible future directions and paves the road ahead.
Open AccessA Survey of Authentication in Internet of Things-Enabled Healthcare Systems
The Internet of medical things (IoMT) provides an ecosystem in which to connect humans, devices, sensors, and systems and improve healthcare services through modern technologies. The IoMT has been around for quite some time, and many architectures/systems have been proposed to exploit its true potential. Healthcare through the Internet of things (IoT) is envisioned to be efficient, accessible, and secure in all possible ways. Even though the personalized health service through IoT is not limited to time or location, many associated challenges have emerged at an exponential pace. With the rapid shift toward IoT-enabled healthcare systems, there is an extensive need to examine possible threats and propose countermeasures. Authentication is one of the key processes in a system’s security, where an individual, device, or another system is validated for its identity. This survey explores authentication techniques proposed for IoT-enabled healthcare systems. The exploration of the literature is categorized with respect to the technology deployment region, as in cloud, fog, and edge. A taxonomy of attacks, comprehensive analysis, and comparison of existing authentication techniques opens up possible future directions and paves the road ahead.
- Open Access
GRA-PIN: A Graphical and PIN-Based Hybrid Authentication Approach for Smart Devices
In many smart devices and numerous digital applications, authentication mechanisms are widely used to validate the legitimacy of users’ identification. As a result of the increased use of mobile devices, most people tend to save sensitive and secret information over such devices. Personal Identification Number (PIN)-based and alphanumeric passwords are simple to remember, but at the same time, they are vulnerable to hackers. Being difficult to guess and more user-friendly, graphical passwords have grown in popularity as an alternative to all such textual passwords. This paper describes an innovative, hybrid, and much more robust user authentication approach, named GRA-PIN (GRAphical and PIN-based), which combines the merits of both graphical and pin-based techniques. The feature of simple arithmetic operations (addition and subtraction) is incorporated in the proposed scheme, through which random passwords are generated for each login attempt. In the study, we have conducted a comparative study between the GRA-PIN scheme with existing PIN-based and pattern-based (swipe-based) authentications approaches using the standard Software Usability Scale (SUS). The usability score of GRA-PIN was analyzed to be as high as 94%, indicating that it is more reliable and user friendly. Furthermore, the security of the proposed scheme was challenged through an experiment wherein three different attackers, having a complete understanding of the proposed scheme, attempted to crack the technique via shoulder surfing, guessing, and camera attack, but they were unsuccessful.
Open AccessGRA-PIN: A Graphical and PIN-Based Hybrid Authentication Approach for Smart Devices
In many smart devices and numerous digital applications, authentication mechanisms are widely used to validate the legitimacy of users’ identification. As a result of the increased use of mobile devices, most people tend to save sensitive and secret information over such devices. Personal Identification Number (PIN)-based and alphanumeric passwords are simple to remember, but at the same time, they are vulnerable to hackers. Being difficult to guess and more user-friendly, graphical passwords have grown in popularity as an alternative to all such textual passwords. This paper describes an innovative, hybrid, and much more robust user authentication approach, named GRA-PIN (GRAphical and PIN-based), which combines the merits of both graphical and pin-based techniques. The feature of simple arithmetic operations (addition and subtraction) is incorporated in the proposed scheme, through which random passwords are generated for each login attempt. In the study, we have conducted a comparative study between the GRA-PIN scheme with existing PIN-based and pattern-based (swipe-based) authentications approaches using the standard Software Usability Scale (SUS). The usability score of GRA-PIN was analyzed to be as high as 94%, indicating that it is more reliable and user friendly. Furthermore, the security of the proposed scheme was challenged through an experiment wherein three different attackers, having a complete understanding of the proposed scheme, attempted to crack the technique via shoulder surfing, guessing, and camera attack, but they were unsuccessful.
- Open Access
- 17 December 2021
- Published by Wiley in Scientific Programming
- Vol. 2021, 1-11
Machine Learning-Based Detection of Spam Emails
Social communication has evolved, with e-mail still being one of the most common communication means, used for both formal and informal ways. With many languages being digitized for the electronic world, the use of English is still abundant. However, various native languages of different regions are emerging gradually. The Urdu language, coming from South Asia, mostly Pakistan, is also getting its pace as a medium for communications used in social media platforms, websites, and emails. With the increased usage of emails, Urdu’s number and variety of spam content also increase. Spam emails are inappropriate and unwanted messages usually sent to breach security. These spam emails include phishing URLs, advertisements, commercial segments, and a large number of indiscriminate recipients. Thus, such content is always a hazard for the user, and many studies have taken place to detect such spam content. However, there is a dire need to detect spam emails, which have content written in Urdu language. The proposed study utilizes the existing machine learning algorithms including Naive Bayes, CNN, SVM, and LSTM to detect and categorize e-mail content. According to our findings, the LSTM model outperforms other models with a highest score of 98.4% accuracy.
- 17 December 2021
- Published by Wiley in Scientific Programming
- Vol. 2021, 1-11
Open AccessMachine Learning-Based Detection of Spam Emails
Social communication has evolved, with e-mail still being one of the most common communication means, used for both formal and informal ways. With many languages being digitized for the electronic world, the use of English is still abundant. However, various native languages of different regions are emerging gradually. The Urdu language, coming from South Asia, mostly Pakistan, is also getting its pace as a medium for communications used in social media platforms, websites, and emails. With the increased usage of emails, Urdu’s number and variety of spam content also increase. Spam emails are inappropriate and unwanted messages usually sent to breach security. These spam emails include phishing URLs, advertisements, commercial segments, and a large number of indiscriminate recipients. Thus, such content is always a hazard for the user, and many studies have taken place to detect such spam content. However, there is a dire need to detect spam emails, which have content written in Urdu language. The proposed study utilizes the existing machine learning algorithms including Naive Bayes, CNN, SVM, and LSTM to detect and categorize e-mail content. According to our findings, the LSTM model outperforms other models with a highest score of 98.4% accuracy.