Research
My research focuses on secure and resilient smart systems, with emphasis on IoT video manipulation attacks, lightweight detection/defense, and interpretable methods for resource-constrained environments.
I am currently pursuing a PhD in Computer Science at the University of Essex, where my research centers on cybersecurity challenges in IoT and smart systems. My work focuses on identifying practical threats and building lightweight, interpretable mechanisms for detection and defense in real-world, resource-constrained environments.
A key focus of my research is video manipulation attacks in IoT-enabled systems, including the design and analysis of detection/defense approaches that are feasible on embedded platforms rather than only in high-resource environments.
My broader research approach combines threat modeling, experimental validation, systems-level analysis, and practical implementation thinking to ensure that proposed methods are not only effective, but also deployable.
Security threats and defensive mechanisms for video-enabled IoT and smart monitoring systems.
A novel video manipulation threat model in which frames are altered dynamically during live operation, with emphasis on timing, context-aware manipulation, and evasion of conventional assumptions.
Designing methods that can run on constrained systems (e.g., edge/embedded devices), balancing detection quality, performance overhead, and implementation practicality.
Using structured experiments and measurable system behavior to evaluate attack impact and defensive response, supporting practical deployment considerations rather than purely theoretical results.
The recurring technical themes that shape my work.
Threat analysis and practical security mechanisms for interconnected smart devices and sensing/monitoring systems.
Security implications of tampered visual streams and the challenges of preserving trust in video-based IoT systems.
Designing solutions that consider CPU, memory, timing, and deployment limitations in constrained environments.
Building mechanisms that are understandable, measurable, and easier to evaluate and trust in real deployments.
A simplified view of how I approach practical cybersecurity research in this area.
Define realistic attack behavior, assumptions, and system context for IoT video-enabled environments.
Build reproducible test scenarios and collect measurable system/stream behavior under normal and attack conditions.
Design lightweight mechanisms suitable for resource-constrained devices and edge deployments.
Assess effectiveness, overhead, interpretability, and practical deployment considerations.
Add your papers, conference presentations, posters, and workshops here as your site grows.
The advent of Internet of Things (IoT) technology has revolutionized disaster management, providing real-time monitoring and enhanced response capabilities for critical situations like floods. Despite these advances, IoT systems are increasingly vulnerable to cyber-attacks, particularly data manipulation attacks that target video feeds. This paper presents a novel attack technique named Dynamic Frame Alternation (DFA) aimed at evading standard detection algorithms. Unlike traditional attacks like replay, frame injection, and video stream hijacking that modify frames in a linear or bulk approach, DFA strategically alters frames based on real-time changes in video attributes, such as color consistency and object presence. Leveraging metrics like the Structural Similarity Index Measure (SSIM) to detect optimal moments for frame manipulation, DFA enhances attack stealth by maintaining low detectability and resource usage. Experimental results, obtained from implementations on an embedded board platform, demonstrate that DFA consistently achieves lower detection rates when compared against traditional attacks while applying popular detection algorithms.
The rapid expansion of IoT video systems in security-critical environments has increased exposure to sophisticated video manipulation attacks. Traditional detection methods including machine learning approaches such as SVM, kNN, and LOF show strong performance against replay, frame injection, and stream hijacking attacks, but fail to detect Dynamic Frame Alteration (DFA), a subtle and adaptive manipulation technique that introduces minimal visual disturbance and leaves limited statistical traces in pixel space. To address this limitation, this work presents a real-time DFA detection framework that relies on system-level behavioral metrics rather than video content analysis. The proposed method monitors frame rate, CPU usage, memory load, and process activity using sliding time windows, rolling averages, and empirically validated thresholds, enabling robust detection. Experimental results show that DFA manipulation produces consistent deviations in these system metrics,particularly frame drops and CPU/memory spikes which are effectively captured by the proposed approach. Across two realistic scenarios the detector achieves up to 93% detection accuracy, significantly outperforming traditional machine learning models whose detection rates on DFA average below 25%.These findings establish system-metric analysis as a practical and explainable alternative to machine learning methods to detect advanced video manipulation in IoT environments.
Areas I’m actively interested in exploring and collaborating on.
I’m open to collaborations across academia and industry, especially in applied cybersecurity, IoT security, smart systems resilience, and practical defense engineering.
If your work aligns with these areas, feel free to reach out to discuss research ideas, joint publications, technical projects, or speaking opportunities.