Programme: Computer Engineering

Now showing 1 - 2 of 2

DEVELOPMENT OF A PAIRING-ANOMALY DETECTION SYSTEM IN VIDEO SURVEILLANCE USING OBJECT DETECTION MODEL
(2025-03) SODIPO, QUEEN BUSAYO; Covenant University Diseertation
Anomaly detection in social behaviours among students is critical to maintaining a safe and respectful academic environment. This research focuses on developing an object detection-based pairing-anomaly detection system to identify and classify unusual social behaviours, such as hand-holding, which can be perceived as an anomaly and potentially lead to more serious issues like sexual misconduct. Object detection, a most significant branch of computer vision, attempts to locate and identify objects in images or video frames. The other is anomaly detection, identifying data sets that do not follow expected behaviour configurations. Combining object detection and anomaly detection approaches provides a powerful solution to detect and flag anomalous or problematic behaviour in any setting. Object detection-based models have recently gained significant attention for their efficiency in detecting and identifying anomalies of interest in complex scenes at high precision. This study shows the application of deep learning for anomalous event detection with an object detection architecture. Utilising the YOLO (You Only Look Once) architecture, the system is designed to detect and localise anomalies in real-time. The model is trained on a custom holding hands dataset, specifically designed to capture instances of hand-holding alongside the Pascal VOC (Visual Object Classes) benchmark dataset, to ensure versatility across varied scenarios. This approach incorporates transfer learning and data augmentation techniques to and optimise model performance with limited labelled data. Evaluation metrics, including mean accuracy precision (mAP), recall, F1 score, and AUC (Area Under Curve), demonstrate the model's effectiveness, with the Custom Holding Hands Dataset achieving an impressive mAP score of 99.5%. The system is integrated into a web application, enabling real-time anomaly detection and classification. This research contributes to developing computer vision-based pairing-anomaly detection systems for social behaviour analysis, with potential applications in maintaining a safe and respectful academic environment.
DEVELOPMENT OF AN AUTONOMOU AGENT FOR A NUMBER STRATEGY GAME USING DEEP Q-NETWORK
(Covenant University Ota, 2025-03) NKWOR, JANE CHINELO; Covenant University Dissertation
Deep Q-Networks (DQNs) have emerged as a pivotal reinforcement learning algorithm for training autonomous agents in complex decision-making tasks. This study investigates the application of Deep Q-Networks in Numero, a number strategy game that requires logical reasoning and iterative feedback processing. Numero is a number strategy game where players predict an opponent's secret four-digit number in the fewest steps possible by analysing feedback and refining strategies. The study explores Numero's unique challenges, such as sparse reward structures, high-dimensional state-action spaces, and non-deterministic feedback mechanisms. To address these challenges, a Deep Q-Network algorithm augmented with Prioritised Experience Replay(PER) was designed and developed to enhance sample efficiency by prioritising critical experiences during training. The autonomous agent interacts with the custom environment, sampling mini-batches from the replay buffer, performing backpropagation, and updating Q-values to improve decision-making. Hyperparameters, such as learning rate, discount factor, replay buffer and exploration rate, were tuned to optimise the agent's learning efficiency. Comparative analysis was conducted using Reservoir Sampling without Replacement and the Minimax algorithm as a baseline approach. Experimental results show that the algorithm achieved a higher success rate (correctly predicted numbers) and faster convergence than Minimax, reducing the average number of steps required to guess the secret number by more than 100%. Additionally, this algorithm demonstrated superior adaptability in handling dynamic feedback, outperforming Reservoir sampling in long-term decision-making. These findings reveal the effectiveness of Deep Q-Networks in structured feedback-driven environments, suggesting their potential application in logical reasoning and decision-making tasks and that the autonomous agent learns effective decision-making strategies through iterative training and fine-tuning, demonstrating improved performance in predicting the opponent's secret number. Further research directions include extending this approach to multi-agent settings where multiple autonomous agents can compete or collaborate to refine their strategic reasoning and explore its application in real-world scenarios requiring structured feedback processing.

Programme: Computer Engineering

News

Browse

Filters

Settings

Sort By

Results per page

Search Results