AI-Enabled Smart Vehicle with CNN and Deep Face for Anti-Theft Detection System

This paper addresses the urgent need for vehicle safety and security amidst the escalating occurrences of road accidents and thefts. It delves into the shortcomings of existing security systems and their ramifications on public safety and property loss. It investigates the fusion of IoT devices and neural networks to reinforce vehicle security measures. It provides a comprehensive analysis of how face detection in real time was achieved using the Haar-Cascade Algorithm and face matching was through the DeepFace technique. Additionally, the implementation includes real-time alerts via email utilizing the Simple Mail Transfer Protocol (SMTP), with attachments showcasing unauthorized individuals. Notably, the integration of CNN enables real-time threat detection, thereby mitigating the risks associated with road accidents and thefts. The accuracy, precision, recall, and F1 scores of CNN applied for DeepFace are mentioned in result section also compared with existing work. This paper concludes by highlighting its findings, emphasizing the efficacy of IoT-based vehicle security systems employing neural networks, which boost both cost and energy efficiency owing to their reduced sensor count.