Niharika Anand
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Niharika Anand has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Stilianos Vidalis
School of Physics, Engineering and Computer Science, University of Hertfordshire, Hatfield AL10 9AB, UK
5 shared publications - Gaurav Sharma
University of Hertfordshire, Hatfield, UK
4 shared publications - Somesh Kumar
Atal Bihari Vajpayee Indian Institute of Information Technology and Management
4 shared publications - Catherine Menon
University of Hertfordshire
3 shared publications - Gaurav Sharma
University of Sheffiled, Sheffiled, England, UK
3 shared publications - Vishal Krishna Singh
University of Essex
3 shared publications
Latest Publications
- 6 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (5) , 8309-8322
Low-Light Image Enhancement for Edge-Based Security Surveillance in 6G-IoT Visual Systems
Application areas such as real-time visual analytics over high-bandwidth 6G networks, low-power camera networks in remote or low-light environments, and surveillance drones usually operate under insufficient lighting conditions. The captured images are often of low quality, poor resolution, and poor visual clarity, leading to reduced visibility, color distortion, and amplified noise. Existing methods of low-light image enhancement (LLIE) suffer from low accuracy with compromised reliability, trust, and fairness. Inspired by the zero-reference learning paradigm of Zero-DCE++, this work aims to investigate the impact of data preprocessing and augmentation strategies for improving the performance of real-time, mission-critical security systems where low-light surveillance images are used for critical decision-making. The proposed method uses fast and flexible denoising convolutional neural network (FFDNet) for denoising, exposure fusion for illumination improvement, and data augmentation for bias mitigation and performance optimization through diverse training samples. The method is curated for edge deployment on constrained IoT hardware, with low latency and energy efficient usage in 6G-IoT visual systems. The proposed model is aimed at performance improvement on trust-driven visual improvements, reduced distributional bias, and deployment fairness across diverse lighting conditions and scenarios. Comparative analysis demonstrates that with the help of zero-reference deep curve estimation, the proposed, DA-Zero-DCE++, pipeline achieves improved performance compared to state-of-the-art LLIE methods. Our best configuration, which combines exposure fusion-based augmentation and mild denoising using FFDNet, achieves an average peak signal-to-noise ratio (PSNR) of 15.34 dB, structural similarity index (SSIM) of 0.4869, and mean absolute error (MAE) of 40.87 on the single image contrast enhancement (SICE) dataset at $1200 \times 900$ resolution. For high-level vision applications such as real-time visual analytics over high-bandwidth 6G networks and low-power camera networks in remote or low-light environments, the performance is further validated on the DarkFace dataset where a high average precision (AP) at intersection over union (IoU) of 0.5 is achieved.
- 6 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (5) , 8309-8322
Low-Light Image Enhancement for Edge-Based Security Surveillance in 6G-IoT Visual Systems
Application areas such as real-time visual analytics over high-bandwidth 6G networks, low-power camera networks in remote or low-light environments, and surveillance drones usually operate under insufficient lighting conditions. The captured images are often of low quality, poor resolution, and poor visual clarity, leading to reduced visibility, color distortion, and amplified noise. Existing methods of low-light image enhancement (LLIE) suffer from low accuracy with compromised reliability, trust, and fairness. Inspired by the zero-reference learning paradigm of Zero-DCE++, this work aims to investigate the impact of data preprocessing and augmentation strategies for improving the performance of real-time, mission-critical security systems where low-light surveillance images are used for critical decision-making. The proposed method uses fast and flexible denoising convolutional neural network (FFDNet) for denoising, exposure fusion for illumination improvement, and data augmentation for bias mitigation and performance optimization through diverse training samples. The method is curated for edge deployment on constrained IoT hardware, with low latency and energy efficient usage in 6G-IoT visual systems. The proposed model is aimed at performance improvement on trust-driven visual improvements, reduced distributional bias, and deployment fairness across diverse lighting conditions and scenarios. Comparative analysis demonstrates that with the help of zero-reference deep curve estimation, the proposed, DA-Zero-DCE++, pipeline achieves improved performance compared to state-of-the-art LLIE methods. Our best configuration, which combines exposure fusion-based augmentation and mild denoising using FFDNet, achieves an average peak signal-to-noise ratio (PSNR) of 15.34 dB, structural similarity index (SSIM) of 0.4869, and mean absolute error (MAE) of 40.87 on the single image contrast enhancement (SICE) dataset at $1200 \times 900$ resolution. For high-level vision applications such as real-time visual analytics over high-bandwidth 6G networks and low-power camera networks in remote or low-light environments, the performance is further validated on the DarkFace dataset where a high average precision (AP) at intersection over union (IoU) of 0.5 is achieved.
- 23 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71 (2) , 5173-5188
Anomaly Detection in Dynamic Graphs Using Multiple Encoding Strategies via Transformers
Accurate anomaly detection in dynamic graph networks suffers due to lack of coverage of all aspects of information; specifically temporal, spatial and centrality based cross-coupled information. This work aims to address the challenge of precise and accurate anomaly detection in dynamic graph networks. It uses a graph-based diffusion technique to sample a fixed-size, yet cross-coupled, information-rich circumstantial node set for target edges. Centrality enabled spatial-temporal node encoding is considered as input to the dynamic graph based transformer network. The proposed method uses a set of four elements to make up the node encoding. The four encoding terms are combined to create an input that contains extensive centrality based cross-coupled spatial-temporal node encoding. The transformer module simultaneously captures all the required attributes with a single encoder. The performance of the proposed method is validated on six different datasets; UCI Messages, Bitcoin-Alpha, Digg Social, Enron Email, Epinions-Trust and AS-Topology. The proposed method outperforms the existing methods in terms of AUC-ROC score, accuracy, loss, and precision. Results show an improvement of 2.42% AUC-ROC value over the existing methods proving the models ability to counter over-fitting and provide accurate results.
- 23 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71 (2) , 5173-5188
Anomaly Detection in Dynamic Graphs Using Multiple Encoding Strategies via Transformers
Accurate anomaly detection in dynamic graph networks suffers due to lack of coverage of all aspects of information; specifically temporal, spatial and centrality based cross-coupled information. This work aims to address the challenge of precise and accurate anomaly detection in dynamic graph networks. It uses a graph-based diffusion technique to sample a fixed-size, yet cross-coupled, information-rich circumstantial node set for target edges. Centrality enabled spatial-temporal node encoding is considered as input to the dynamic graph based transformer network. The proposed method uses a set of four elements to make up the node encoding. The four encoding terms are combined to create an input that contains extensive centrality based cross-coupled spatial-temporal node encoding. The transformer module simultaneously captures all the required attributes with a single encoder. The performance of the proposed method is validated on six different datasets; UCI Messages, Bitcoin-Alpha, Digg Social, Enron Email, Epinions-Trust and AS-Topology. The proposed method outperforms the existing methods in terms of AUC-ROC score, accuracy, loss, and precision. Results show an improvement of 2.42% AUC-ROC value over the existing methods proving the models ability to counter over-fitting and provide accurate results.
- 2 July 2024
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 80 (16) , 24559-24577
Convolutional neural network (CNN) and federated learning-based privacy preserving approach for skin disease classification
- 2 July 2024
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 80 (16) , 24559-24577
Convolutional neural network (CNN) and federated learning-based privacy preserving approach for skin disease classification
- 30 May 2024
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 84 (13) , 12241-12264
Unleashing the power of Devops and Iot: a framework for software delivery and system availability
- 30 May 2024
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 84 (13) , 12241-12264
Unleashing the power of Devops and Iot: a framework for software delivery and system availability
- 13 May 2024
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 84 (12) , 10109-10136
Automated passive income from stock market using machine learning and big data analytics with security aspects
- 13 May 2024
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 84 (12) , 10109-10136
Automated passive income from stock market using machine learning and big data analytics with security aspects
- Open Access
- 31 December 2023
- Published by International Consortium of Academic Professionals for Scientific Research in International Journal for Global Academic & Scientific Research
- Vol. 2 (4) , 02
Automated Passive Income from Stock Market Using Machine Learning and Big Data Analytics with Security Aspects-Retracted
This article is retracted: Dear IJGASR Readers, We regret to inform you that the article titled “Automated Passive Income from Stock Market Using Machine Learning and Big Data Analytics with Security Aspects,” published in Volume 2, Issue 4, 2023, with DOI https://doi.org/10.55938/ijgasr.v2i4.63 has been retracted. The retraction is based on the request of the author, Dr. Gaurav Sharma, as it he along with his team is willing to expand his research further & want to publish with more research findings. After thorough examination and review of the article, Editorial team has decided to consider request for the retraction/withdrawal of article We apologize for any inconvenience or confusion this may have caused and reaffirm our commitment to upholding the integrity of our publication. We thank the author for bringing this matter to our attention and for taking the appropriate steps to retract the article. Sincerely, Editorial Team, IJGASR Announcement: https://journals.icapsr.com/index.php/ijgasr/announcement/view/27
- 31 December 2023
- Published by International Consortium of Academic Professionals for Scientific Research in International Journal for Global Academic & Scientific Research
- Vol. 2 (4) , 02
Open AccessAutomated Passive Income from Stock Market Using Machine Learning and Big Data Analytics with Security Aspects-Retracted
This article is retracted: Dear IJGASR Readers, We regret to inform you that the article titled “Automated Passive Income from Stock Market Using Machine Learning and Big Data Analytics with Security Aspects,” published in Volume 2, Issue 4, 2023, with DOI https://doi.org/10.55938/ijgasr.v2i4.63 has been retracted. The retraction is based on the request of the author, Dr. Gaurav Sharma, as it he along with his team is willing to expand his research further & want to publish with more research findings. After thorough examination and review of the article, Editorial team has decided to consider request for the retraction/withdrawal of article We apologize for any inconvenience or confusion this may have caused and reaffirm our commitment to upholding the integrity of our publication. We thank the author for bringing this matter to our attention and for taking the appropriate steps to retract the article. Sincerely, Editorial Team, IJGASR Announcement: https://journals.icapsr.com/index.php/ijgasr/announcement/view/27
- Open Access
- 5 October 2023
- Published by Springer Nature in Research Square
Exponentiating the Precision of Facial Spoof Detection: An Intricately Woven Tapestry of Hybrid Machine Learning Paradigms and State-of-the-Art Textural Feature Extraction Methodologies
In recent times, there has been a rapid increase in the deployment of face recognition systems,creating a significant demand for stringent safety regulations to govern their implementation. Several approaches have been developed for face spoof detection, utilizing techniques such as feature extraction, machine learning (ML) and deep learning (DL). These methods have gained significant interest in various domains due to their accessibility and ability to capture abundant information in daily life. Face spoof detection techniques based on ML and Artificial Intelligence (AI) typically involve multiple stages, including pre-processing the raw data, feature extraction, and face classification. Previous methods have employed K-Nearest Neighbor (KNN) in combination with Grey Level Co-occurrence Matrix (GLCM) for detecting face spoofing, but with relatively low accuracy.In this study, a novel approach is introduced that combines Histogram of Oriented Gradients (HOG) with GLCM for feature extraction. These features are then utilized in a hybrid classifier, which integrates the Random Forest (RF) algorithm boosted with Adaboost optimizer, K-Nearest Neighbor (KNN), and XGBoost as classification algorithms. Through extensive evaluation of our proposed strategies and thorough comparison, our approach demonstrates its superiority over alternative methods, achieving remarkable levels of accuracy, precision, and recall performance.
- 5 October 2023
- Published by Springer Nature in Research Square
Open AccessExponentiating the Precision of Facial Spoof Detection: An Intricately Woven Tapestry of Hybrid Machine Learning Paradigms and State-of-the-Art Textural Feature Extraction Methodologies
In recent times, there has been a rapid increase in the deployment of face recognition systems,creating a significant demand for stringent safety regulations to govern their implementation. Several approaches have been developed for face spoof detection, utilizing techniques such as feature extraction, machine learning (ML) and deep learning (DL). These methods have gained significant interest in various domains due to their accessibility and ability to capture abundant information in daily life. Face spoof detection techniques based on ML and Artificial Intelligence (AI) typically involve multiple stages, including pre-processing the raw data, feature extraction, and face classification. Previous methods have employed K-Nearest Neighbor (KNN) in combination with Grey Level Co-occurrence Matrix (GLCM) for detecting face spoofing, but with relatively low accuracy.In this study, a novel approach is introduced that combines Histogram of Oriented Gradients (HOG) with GLCM for feature extraction. These features are then utilized in a hybrid classifier, which integrates the Random Forest (RF) algorithm boosted with Adaboost optimizer, K-Nearest Neighbor (KNN), and XGBoost as classification algorithms. Through extensive evaluation of our proposed strategies and thorough comparison, our approach demonstrates its superiority over alternative methods, achieving remarkable levels of accuracy, precision, and recall performance.
- 30 September 2023
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 80 (4) , 5540-5564
Applications hosting over cloud-assisted IOT: a productivity model and method defining accessibility of data security
- 30 September 2023
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 80 (4) , 5540-5564
Applications hosting over cloud-assisted IOT: a productivity model and method defining accessibility of data security
- 2 November 2022
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 82 (11) , 16951-16971
Analysis and implementation of semi-automatic model for vulnerability exploitations of threat agents in NIST databases
Proactive security plays a vital role in preventing the attack before entering active mode. In the modern information environment, it depends on the vulnerability management practitioners of an organization in which the critical factor is the prioritization of threats. The existing models and methodology follow the traditional approaches of a Common Vulnerability Scoring System (CVSS) to prioritize threats and vulnerabilities. The CVSS is not able to provide effectiveness to the security of the business of an organization. In contrast, the vulnerability analysis needs a model which can give significance to the prioritization policies. The model depends on the CVSS score of threats and compares various features of vulnerability like threat vectors, inputs, environments used by threat agent’s groups, and potential outputs of threat agents. Therefore, the research aims to design a semi-automatic model for vulnerability analysis of threats for the National Institute of Standards and Technology (NIST) database of cyber-crime. We have developed a semi-automatic model that simulates the CVE (Common Vulnerabilities and Exposures) list of the NIST database between 1999 and 2021, concerning the resources used by the threat agents, pre-requisites input, attack vectors, and dormant results. The semi-automatic approach of the model to perform the vulnerability analysis of threat agent groups identified in a network makes the model more efficient and effective to addresses the profiling of threat agents and evaluating the CTI (Critical Threat intelligence feed). Our experimental results imply that the semi-automatic model implements the vulnerability prioritization based on the CVSS score and uses the comparative analysis based on the threat agent’s vectors identified. It also provides potency and optimized complexity to an organization’s business to mitigate the vulnerability identified in a network.
- 2 November 2022
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 82 (11) , 16951-16971
Analysis and implementation of semi-automatic model for vulnerability exploitations of threat agents in NIST databases
Proactive security plays a vital role in preventing the attack before entering active mode. In the modern information environment, it depends on the vulnerability management practitioners of an organization in which the critical factor is the prioritization of threats. The existing models and methodology follow the traditional approaches of a Common Vulnerability Scoring System (CVSS) to prioritize threats and vulnerabilities. The CVSS is not able to provide effectiveness to the security of the business of an organization. In contrast, the vulnerability analysis needs a model which can give significance to the prioritization policies. The model depends on the CVSS score of threats and compares various features of vulnerability like threat vectors, inputs, environments used by threat agent’s groups, and potential outputs of threat agents. Therefore, the research aims to design a semi-automatic model for vulnerability analysis of threats for the National Institute of Standards and Technology (NIST) database of cyber-crime. We have developed a semi-automatic model that simulates the CVE (Common Vulnerabilities and Exposures) list of the NIST database between 1999 and 2021, concerning the resources used by the threat agents, pre-requisites input, attack vectors, and dormant results. The semi-automatic approach of the model to perform the vulnerability analysis of threat agent groups identified in a network makes the model more efficient and effective to addresses the profiling of threat agents and evaluating the CTI (Critical Threat intelligence feed). Our experimental results imply that the semi-automatic model implements the vulnerability prioritization based on the CVSS score and uses the comparative analysis based on the threat agent’s vectors identified. It also provides potency and optimized complexity to an organization’s business to mitigate the vulnerability identified in a network.
- Open Access
- 28 September 2021
- Published by MDPI AG in Electronics
- Vol. 10 (19) , 2365
A Survey on Layer-Wise Security Attacks in IoT: Attacks, Countermeasures, and Open-Issues
Security is a mandatory issue in any network, where sensitive data are transferred safely in the required direction. Wireless sensor networks (WSNs) are the networks formed in hostile areas for different applications. Whatever the application, the WSNs must gather a large amount of sensitive data and send them to an authorized body, generally a sink. WSN has integrated with Internet-of-Things (IoT) via internet access in sensor nodes along with internet-connected devices. The data gathered with IoT are enormous, which are eventually collected by WSN over the Internet. Due to several resource constraints, it is challenging to design a secure sensor network, and for a secure IoT it is essential to have a secure WSN. Most of the traditional security techniques do not work well for WSN. The merger of IoT and WSN has opened new challenges in designing a secure network. In this paper, we have discussed the challenges of creating a secure WSN. This research reviews the layer-wise security protocols for WSN and IoT in the literature. There are several issues and challenges for a secure WSN and IoT, which we have addressed in this research. This research pinpoints the new research opportunities in the security issues of both WSN and IoT. This survey climaxes in abstruse psychoanalysis of the network layer attacks. Finally, various attacks on the network using Cooja, a simulator of ContikiOS, are simulated.
- 28 September 2021
- Published by MDPI AG in Electronics
- Vol. 10 (19) , 2365
Open AccessA Survey on Layer-Wise Security Attacks in IoT: Attacks, Countermeasures, and Open-Issues
Security is a mandatory issue in any network, where sensitive data are transferred safely in the required direction. Wireless sensor networks (WSNs) are the networks formed in hostile areas for different applications. Whatever the application, the WSNs must gather a large amount of sensitive data and send them to an authorized body, generally a sink. WSN has integrated with Internet-of-Things (IoT) via internet access in sensor nodes along with internet-connected devices. The data gathered with IoT are enormous, which are eventually collected by WSN over the Internet. Due to several resource constraints, it is challenging to design a secure sensor network, and for a secure IoT it is essential to have a secure WSN. Most of the traditional security techniques do not work well for WSN. The merger of IoT and WSN has opened new challenges in designing a secure network. In this paper, we have discussed the challenges of creating a secure WSN. This research reviews the layer-wise security protocols for WSN and IoT in the literature. There are several issues and challenges for a secure WSN and IoT, which we have addressed in this research. This research pinpoints the new research opportunities in the security issues of both WSN and IoT. This survey climaxes in abstruse psychoanalysis of the network layer attacks. Finally, various attacks on the network using Cooja, a simulator of ContikiOS, are simulated.