Object Detection In Real-World Scenarios Using Artificial Intelligence And Machine Learning Technologies

Abstract

This study explores the application of Artificial Intelligence (AI) and Machine Learning (ML) techniques for object detection in real-world scenarios, with a particular focus on Albania. The purpose of the research is to develop and evaluate advanced object detection models that can enhance accuracy and reliability in various applications such as security and road safety. The study employs an experimental approach, leveraging the YOLO algorithm and Convolutional Neural Networks (CNNs) to train and evaluate customized object detection models using diverse datasets. Comparative analyses are conducted to identify the most effective methodologies. The findings demonstrate that larger, high-quality datasets significantly enhance model performance, as evidenced by a maximum F1-score of 0.96 achieved with 80 training images and 50 epochs. The research highlights the transformative potential of AI-driven object detection in improving processing speed and accuracy for critical applications. Challenges such as computational resource limitations and dataset constraints are identified as barriers to broader implementation. The study concludes with practical recommendations for improving model scalability and reliability, emphasizing the importance of integrating AI with complementary technologies for real-world deployment. These insights have implications for policymakers, developers, and industries aiming to leverage AI for enhanced safety and efficiency in infrastructure and beyond.