Machine Learning Approaches for Abandoned Luggage Detection

Security of public spaces is a major concern, and one of the critical challenges in this context is the timely detection of abandoned luggage, which may pose security threats. Traditional methods for abandoned luggage detection often rely on manual monitoring, making them resource-intensive and prone to human error. In this paper, we present a comprehensive study on the application of machine learning techniques for abandoned luggage detection.Our research work focuses on utilization of Convolutional Neural Networks (CNNs) as well as conventional machine learning algorithms, involving Support Vector Machines (SVMs) and ensemble methods. We describe the data collection and preprocessing steps, encompassing diverse data sources, including video footage, images, and sensor data. The methodology section outlines the machine learning approaches, feature extraction, model architectures, and hyperparameters selected for the task.We evaluate performance of these models using a range of metrics, including precision, recall, F1-score, and, where applicable, Average Precision (AP). Our results demonstrate that the CNN-based model achieved impressive precision (0.92) and recall (0.87), indicating its effectiveness in accurately identifying abandoned luggage. Ensemble methods, such as Random Forest and Gradient Boosting, also exhibited competitive performance, offering flexibility in balancing precision and recall.