Alamgir Naushad
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Alamgir Naushad has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Abd Ullah Khan
Kyung Hee University, Giheung-gu
3 shared publications - Muhammad Tanveer
Center of Artificial Intelligence, Prince Mohammad Bin Fahd University
3 shared publications - Sibghat Ullah Bazai
Department of Computer Engineering, Balochistan University of Information Technology, Engineering, and Management Sciences (BUITEMS), Quetta 87300, Pakistan
3 shared publications - Shehzad Ashraf Chaudhry
Abu Dhabi University
2 shared publications - Abdullah G. Alharbi
Princess Nourah Bint Abdulrahman University
1 shared publication - Afsheen Sabir
Balochistan University of Information Technology, Engineering, and Management Sciences, Pakistan
1 shared publication
Latest Publications
- Open Access
- 23 September 2026
- Published by PeerJ in PeerJ Computer Science
- Vol. 12, e4074
Robust phishing URL detection using FastText embeddings and deep sequential models
Phishing attacks are increasingly prevalent, leading to the loss of assets, financial information, personal data, and other sensitive information to malicious actors. Various detection methods have been developed to protect users from such attacks. Traditional phishing Uniform Resource Locator (URL) detection methods, such as blacklisting and heuristics, rely on maintaining lists of known phishing URLs. However, these techniques struggle to identify new unlisted URLs and are expensive to maintain. In recent years, there has been a shift toward deep learning techniques, which are particularly efficient for training large data systems and handling undefined features. This research aims to develop and compare seven deep-learning models to identify phishing URLs more accurately and efficiently. The proposed model utilizes FastText-based word embeddings to preprocess input URLs, generating a robust vector representation. This model is then applied to various deep learning-based classification models, including Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) network, Bidirectional LSTM (BiLSTM), CNN-LSTM hybrid model, CNN-BiLSTM hybrid model, Gated Recurrent Unit (GRU), and Bidirectional GRU (BiGRU). The experimental results demonstrate that the BiLSTM model with FastText embeddings outperforms alternative models, attaining an accuracy of 98.62%. The proposed technique was tested on benchmark data from an online repository and then validated on an independent external dataset to assess its resilience and generalization capabilities, yielding good phishing-detection performance across multiple data sources. Even though some earlier research has achieved an accuracy above 99%, many of these studies ignore token-level URL structures, avoid hybrid architectures, or depend on small or unbalanced datasets. To overcome these constraints, we combine FastText embeddings with BiLSTM networks to improve generalization and robustness, especially for identifying obfuscated and unique phishing URLs. This results in a more dependable and scalable detection system.
- 23 September 2026
- Published by PeerJ in PeerJ Computer Science
- Vol. 12, e4074
Open AccessRobust phishing URL detection using FastText embeddings and deep sequential models
Phishing attacks are increasingly prevalent, leading to the loss of assets, financial information, personal data, and other sensitive information to malicious actors. Various detection methods have been developed to protect users from such attacks. Traditional phishing Uniform Resource Locator (URL) detection methods, such as blacklisting and heuristics, rely on maintaining lists of known phishing URLs. However, these techniques struggle to identify new unlisted URLs and are expensive to maintain. In recent years, there has been a shift toward deep learning techniques, which are particularly efficient for training large data systems and handling undefined features. This research aims to develop and compare seven deep-learning models to identify phishing URLs more accurately and efficiently. The proposed model utilizes FastText-based word embeddings to preprocess input URLs, generating a robust vector representation. This model is then applied to various deep learning-based classification models, including Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) network, Bidirectional LSTM (BiLSTM), CNN-LSTM hybrid model, CNN-BiLSTM hybrid model, Gated Recurrent Unit (GRU), and Bidirectional GRU (BiGRU). The experimental results demonstrate that the BiLSTM model with FastText embeddings outperforms alternative models, attaining an accuracy of 98.62%. The proposed technique was tested on benchmark data from an online repository and then validated on an independent external dataset to assess its resilience and generalization capabilities, yielding good phishing-detection performance across multiple data sources. Even though some earlier research has achieved an accuracy above 99%, many of these studies ignore token-level URL structures, avoid hybrid architectures, or depend on small or unbalanced datasets. To overcome these constraints, we combine FastText embeddings with BiLSTM networks to improve generalization and robustness, especially for identifying obfuscated and unique phishing URLs. This results in a more dependable and scalable detection system.
- 14 May 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Pattern Recognition and Artificial Intelligence
- Vol. 40 (12)
Facial Expression and Key Landmarks Detection Using Deep Learning Techniques
The recognition of facial expressions is a crucial topic in the field of computer vision and artificial intelligence due to its significant contribution in providing an essential understanding of human emotions and motivations. However, accurately detecting human emotions through algorithms poses a significant challenge because facial expressions are interpreted as mathematical patterns. The training capabilities of conventional machine learning techniques are limited, leading to reduced rates of recognition. This study focuses on utilizing transfer learning methodologies to enhance emotion identification. Notably, the authors achieved significant accuracy rates by utilizing pre-trained ResNet-50, VGG-19, and Inception v3 networks. The ResNet-50 model achieved a 98% accuracy rate, while the VGG-19 and Inception v3 models attained 96% and 92% accuracy rates, respectively. The use of pre-trained models with their corresponding pre-trained weights resulted in improved accuracy of facial expression recognition (FER). Furthermore, the study results were compared to conventional machine learning methods of the past, and the proposed transfer learning approach exhibited enhanced precision in identifying facial expressions. The findings of the study suggest that the utilization of transfer learning techniques can significantly improve the accuracy of models designed for recognizing facial expressions. FER holds significant promise in a variety of domains such as emotion analysis, human–computer interaction, and security systems. This investigation highlights the potential of transfer learning methodologies in advancing the field of FER, thereby facilitating a better comprehension of human emotion and motivation.
- 14 May 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Pattern Recognition and Artificial Intelligence
- Vol. 40 (12)
Facial Expression and Key Landmarks Detection Using Deep Learning Techniques
The recognition of facial expressions is a crucial topic in the field of computer vision and artificial intelligence due to its significant contribution in providing an essential understanding of human emotions and motivations. However, accurately detecting human emotions through algorithms poses a significant challenge because facial expressions are interpreted as mathematical patterns. The training capabilities of conventional machine learning techniques are limited, leading to reduced rates of recognition. This study focuses on utilizing transfer learning methodologies to enhance emotion identification. Notably, the authors achieved significant accuracy rates by utilizing pre-trained ResNet-50, VGG-19, and Inception v3 networks. The ResNet-50 model achieved a 98% accuracy rate, while the VGG-19 and Inception v3 models attained 96% and 92% accuracy rates, respectively. The use of pre-trained models with their corresponding pre-trained weights resulted in improved accuracy of facial expression recognition (FER). Furthermore, the study results were compared to conventional machine learning methods of the past, and the proposed transfer learning approach exhibited enhanced precision in identifying facial expressions. The findings of the study suggest that the utilization of transfer learning techniques can significantly improve the accuracy of models designed for recognizing facial expressions. FER holds significant promise in a variety of domains such as emotion analysis, human–computer interaction, and security systems. This investigation highlights the potential of transfer learning methodologies in advancing the field of FER, thereby facilitating a better comprehension of human emotion and motivation.
- 18 July 2025
- Published by IGI Global Scientific Publishing
- p. 119-164
Big Data Analytics and AI-Driven Approaches for Air Pollution Trend Analysis
To enhance the air quality monitoring and decision-making, this research investigates the combination of Big Data technologies, Machine learning, Deep learning models, and Artificial Intelligence. Using IOT sensor networks, satellite imaging, and meteorological data will be beneficial to meet the volume, velocity, and variety of big data characteristics. Advanced models like CNNs, LSTMs, and CNN-LSTM hybrid models are employed for accurate and precise spatiotemporal forecasting. Such models are evaluated using metrics such as RMSE, MAE, and R2. Domain adaptation and transfer learning are used in regions where there is limited data. AI-powered public health analysis, wearable sensors for real-time exposure and pollution detection in various countries like China and India, are some real-world applications. Despite the limitations in the model's performance and data dependability, this research outlines a future path integrating ethical AI, citizen science, and integration with smart cities for sustainable pollution management.
- 18 July 2025
- Published by IGI Global Scientific Publishing
- p. 119-164
Big Data Analytics and AI-Driven Approaches for Air Pollution Trend Analysis
To enhance the air quality monitoring and decision-making, this research investigates the combination of Big Data technologies, Machine learning, Deep learning models, and Artificial Intelligence. Using IOT sensor networks, satellite imaging, and meteorological data will be beneficial to meet the volume, velocity, and variety of big data characteristics. Advanced models like CNNs, LSTMs, and CNN-LSTM hybrid models are employed for accurate and precise spatiotemporal forecasting. Such models are evaluated using metrics such as RMSE, MAE, and R2. Domain adaptation and transfer learning are used in regions where there is limited data. AI-powered public health analysis, wearable sensors for real-time exposure and pollution detection in various countries like China and India, are some real-world applications. Despite the limitations in the model's performance and data dependability, this research outlines a future path integrating ethical AI, citizen science, and integration with smart cities for sustainable pollution management.
- 18 July 2025
- Published by IGI Global Scientific Publishing
- p. 65-98
Analytics for Health Resilience
This chapter examines how big data analytics transforms the management of chronic disease during natural disasters. It shows how live research from wearable gadgets, electronic health records (EHRs) and social forums can provide early warning, remote monitoring, and smart resource deployment. Example case studies from catastrophes across the world – the 2005 Hurricane Katrina and the Haiti earthquake – and present COVID-19 pandemic prove the value of AI, IoT, and telemedicine for continuity of care. Some key obstacles like data privacy, infrastructure constraints, interoperability are also reviewed. Ethical frameworks, capacity building and technology integration that build on disaster-resilient healthcare systems are highlighted by the chapter. Eventually, it calls for stakeholder cooperation in order to develop equitable, data-based solutions for vulnerable populations in the event of an emergency.
- 18 July 2025
- Published by IGI Global Scientific Publishing
- p. 65-98
Analytics for Health Resilience
This chapter examines how big data analytics transforms the management of chronic disease during natural disasters. It shows how live research from wearable gadgets, electronic health records (EHRs) and social forums can provide early warning, remote monitoring, and smart resource deployment. Example case studies from catastrophes across the world – the 2005 Hurricane Katrina and the Haiti earthquake – and present COVID-19 pandemic prove the value of AI, IoT, and telemedicine for continuity of care. Some key obstacles like data privacy, infrastructure constraints, interoperability are also reviewed. Ethical frameworks, capacity building and technology integration that build on disaster-resilient healthcare systems are highlighted by the chapter. Eventually, it calls for stakeholder cooperation in order to develop equitable, data-based solutions for vulnerable populations in the event of an emergency.
- Open Access
- 10 June 2022
- Published by Telecommunications Association Inc. in Journal of Telecommunications and the Digital Economy
- Vol. 10 (2) , 267-287
Study on Fixed and Dynamic Spectrum Access Models for Cellular Networks
Radio frequency spectrum is a scarce resource, hence its effective and efficient utilization for contemporary and future technologies is of paramount importance. This paper mainly focuses on analysing conventional models for spectrum management, such as command and control and market-based models, and their inefficacy to serve the upcoming technological demands globally. Generally outlined observations concerning spectrum mismanagement include lesser availability of unassigned spectrum and under-utilization of spectrum allocated to passive users. Due to the considerable urge for flexible spectrum assignment framework models and policies, new spectrum assignment approaches, such as Dynamic Spectrum Access, are investigated. Moreover, models such as Licensed Shared Access (LSA), which evolved from Europe, and Citizen Bands Radio Service (CBRS) from North America are also evaluated in this paper as potential choices for future spectrum management in Pakistan. In the aforesaid models, spectrum is shared among multiple users as per time, place and dimensions, keeping in view the security and priority of incumbent and licensed users. In addition, the efficiency of spectrum utilization and economic advantages of these models have also been analysed. Lastly, LSA and CBRS are studied comparatively, and spectrum management suggestions are made for effective implementation in Pakistan.
- 10 June 2022
- Published by Telecommunications Association Inc. in Journal of Telecommunications and the Digital Economy
- Vol. 10 (2) , 267-287
Open AccessStudy on Fixed and Dynamic Spectrum Access Models for Cellular Networks
Radio frequency spectrum is a scarce resource, hence its effective and efficient utilization for contemporary and future technologies is of paramount importance. This paper mainly focuses on analysing conventional models for spectrum management, such as command and control and market-based models, and their inefficacy to serve the upcoming technological demands globally. Generally outlined observations concerning spectrum mismanagement include lesser availability of unassigned spectrum and under-utilization of spectrum allocated to passive users. Due to the considerable urge for flexible spectrum assignment framework models and policies, new spectrum assignment approaches, such as Dynamic Spectrum Access, are investigated. Moreover, models such as Licensed Shared Access (LSA), which evolved from Europe, and Citizen Bands Radio Service (CBRS) from North America are also evaluated in this paper as potential choices for future spectrum management in Pakistan. In the aforesaid models, spectrum is shared among multiple users as per time, place and dimensions, keeping in view the security and priority of incumbent and licensed users. In addition, the efficiency of spectrum utilization and economic advantages of these models have also been analysed. Lastly, LSA and CBRS are studied comparatively, and spectrum management suggestions are made for effective implementation in Pakistan.
- Open Access
- 7 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 19836-19851
RUAM-IoD: A Robust User Authentication Mechanism for the Internet of Drones
The revolutionary advancement in the capabilities of hardware tools, software packages, and communication techniques gave rise to the Internet of Things-supported drone networks (IoD), thereby enabling smooth communication among devices and applications, and impacting drastically the various aspects of human lives. However, with the increasing sophistication in the infrastructure of IoD, new security threats arise that require novel algorithms and schemes as solutions. To this end, several schemes have recently been proposed. However, some schemes cannot perfectly address the novel security aspects associated with IoD environments, while others cannot provide computational or communication efficiency. Motivated by these research gaps in the existing literature, we leverage elliptic curve cryptography along with symmetric encryption and hash function, and propose a novel and robust user authentication mechanism for the IoD, called RUAM-IoD.We validate the security of the established SK formally through the random oracle model. Similarly, we provide informal security analysis to demonstrate the security capabilities of RUAM-IoD against different pernicious security attacks. Likewise, we establish a comparison of the RUAM-IoD with several state-of-the-art authentication schemes to show that RUAM-IoD acquires less storage, communication, and computational cost.
- 7 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 19836-19851
Open AccessRUAM-IoD: A Robust User Authentication Mechanism for the Internet of Drones
The revolutionary advancement in the capabilities of hardware tools, software packages, and communication techniques gave rise to the Internet of Things-supported drone networks (IoD), thereby enabling smooth communication among devices and applications, and impacting drastically the various aspects of human lives. However, with the increasing sophistication in the infrastructure of IoD, new security threats arise that require novel algorithms and schemes as solutions. To this end, several schemes have recently been proposed. However, some schemes cannot perfectly address the novel security aspects associated with IoD environments, while others cannot provide computational or communication efficiency. Motivated by these research gaps in the existing literature, we leverage elliptic curve cryptography along with symmetric encryption and hash function, and propose a novel and robust user authentication mechanism for the IoD, called RUAM-IoD.We validate the security of the established SK formally through the random oracle model. Similarly, we provide informal security analysis to demonstrate the security capabilities of RUAM-IoD against different pernicious security attacks. Likewise, we establish a comparison of the RUAM-IoD with several state-of-the-art authentication schemes to show that RUAM-IoD acquires less storage, communication, and computational cost.
- Open Access
- 26 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 145683-145698
PASKE-IoD: Privacy-Protecting Authenticated Key Establishment for Internet of Drones
Unmanned aerial vehicles/drones are considered an essential ingredient of traffic motoring systems in smart cities. Interconnected drones, also called the Internet of Drones (IoD), gather critical data from the environmental area of interest and transmit the data to a server located at the control room for further processing. This transmission occurs via wireless communication channels, which are exposed to various security risks. Besides this, an External User (EU) occasionally demands access to real-time information stored at a specific drone rather than retrieving data from the server, which requires an efficient Authenticated Session Key Establishment (ASKE) approach to ensure a reliable communication in IoD environment. In this article, we present a Privacy-Protecting ASKE scheme for IoD (PASKE-IoD). PASKE-IoD utilizes Authenticated Encryption (AE) primitive “ASCON,” and hash function “ASCON-hash,” to accomplish the ASKE phase. PASKE-IoD checks the EU’s authenticity before allowing him to access the IoD environment resources. Moreover, PASKE-IoD enables EUs and drones to communicate securely after establishing a session key. Meticulous informal security analysis and security verification are carried out using Scyther to demonstrate that PASKE-IoD is immune to numerous covert security attacks. In addition, Burrows-Abadi-Needham logic is utilized to corroborate the logical exactitude of PASKE-IoD. A comparative analysis is presented to illustrate that PASKE-IoD is efficient and renders more security features than the eminent ASKE scheme.
- 26 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 145683-145698
Open AccessPASKE-IoD: Privacy-Protecting Authenticated Key Establishment for Internet of Drones
Unmanned aerial vehicles/drones are considered an essential ingredient of traffic motoring systems in smart cities. Interconnected drones, also called the Internet of Drones (IoD), gather critical data from the environmental area of interest and transmit the data to a server located at the control room for further processing. This transmission occurs via wireless communication channels, which are exposed to various security risks. Besides this, an External User (EU) occasionally demands access to real-time information stored at a specific drone rather than retrieving data from the server, which requires an efficient Authenticated Session Key Establishment (ASKE) approach to ensure a reliable communication in IoD environment. In this article, we present a Privacy-Protecting ASKE scheme for IoD (PASKE-IoD). PASKE-IoD utilizes Authenticated Encryption (AE) primitive “ASCON,” and hash function “ASCON-hash,” to accomplish the ASKE phase. PASKE-IoD checks the EU’s authenticity before allowing him to access the IoD environment resources. Moreover, PASKE-IoD enables EUs and drones to communicate securely after establishing a session key. Meticulous informal security analysis and security verification are carried out using Scyther to demonstrate that PASKE-IoD is immune to numerous covert security attacks. In addition, Burrows-Abadi-Needham logic is utilized to corroborate the logical exactitude of PASKE-IoD. A comparative analysis is presented to illustrate that PASKE-IoD is efficient and renders more security features than the eminent ASKE scheme.
- 17 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 9 (9) , 6855-6865
A Robust Access Control Protocol for the Smart Grid Systems
Lightweight Cryptography (LWC) based Authenticated Encryption with Associative Data (AEAD) cryptographic primitives require fewer computational and energy resources than conventional cryptographic primitives as a single operation of an AEAD scheme provides confidentiality, integrity, and authenticity of data. This feature of AEAD schemes helps design an Access Control (AC) protocol to be leveraged for enhancing the security of resource-constrained IoT-enabled Smart Grid (SG) system with low computational overhead and fewer cryptographic operations. This paper presents a novel and robust AC protocol called RACP-SG, which aims to enhance the security of resource-constrained IoT-enabled SG systems. RACP-SG employs an LWC-based AEAD scheme, ASCON and the hash function, ASCON-hash, along with elliptic curve cryptography to accomplish the AC phase. Besides, RACP-SG enables a Smart Meter (SM) and a Service Provider (SEP) to mutually authenticate each other and establish a session key while communicating across the public communication channel. By using the session key, the SM can securely transfer the gathered data to the SEP. We verify the security of the session key using the widely accepted random oracle model. Moreover, we conduct Scyther-based and informal security analyses to demonstrate that RACP-SG is protected against various covert security risks, such as replay, impersonation, and de-synchronization attacks. Besides, we present a comparative study to illustrate that RACP-SG renders superior security features while reducing energy, storage, communication, and computational overheads compared to the state-of-the-art.
- 17 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 9 (9) , 6855-6865
A Robust Access Control Protocol for the Smart Grid Systems
Lightweight Cryptography (LWC) based Authenticated Encryption with Associative Data (AEAD) cryptographic primitives require fewer computational and energy resources than conventional cryptographic primitives as a single operation of an AEAD scheme provides confidentiality, integrity, and authenticity of data. This feature of AEAD schemes helps design an Access Control (AC) protocol to be leveraged for enhancing the security of resource-constrained IoT-enabled Smart Grid (SG) system with low computational overhead and fewer cryptographic operations. This paper presents a novel and robust AC protocol called RACP-SG, which aims to enhance the security of resource-constrained IoT-enabled SG systems. RACP-SG employs an LWC-based AEAD scheme, ASCON and the hash function, ASCON-hash, along with elliptic curve cryptography to accomplish the AC phase. Besides, RACP-SG enables a Smart Meter (SM) and a Service Provider (SEP) to mutually authenticate each other and establish a session key while communicating across the public communication channel. By using the session key, the SM can securely transfer the gathered data to the SEP. We verify the security of the session key using the widely accepted random oracle model. Moreover, we conduct Scyther-based and informal security analyses to demonstrate that RACP-SG is protected against various covert security risks, such as replay, impersonation, and de-synchronization attacks. Besides, we present a comparative study to illustrate that RACP-SG renders superior security features while reducing energy, storage, communication, and computational overheads compared to the state-of-the-art.