An intelligentintegrated vehicle surveillance system for controlling the vehicle thefts/hacking using IoT and facial recognition

Intelligent Integrated Vehicle Surveillance System (IIVSS) is one of the novel techniques developed by the researcher to detect and validate the authorized driver of the vehicle. IIVSS is a machine learning (ML) model built on Python that is intended to accurately recognize the driver's face. By implementing the application there is no requirement for keys, which previously used key card entry for unlocking the car and starting the engine. In addition to the features listed above, the authors have created a DDS (Drowsiness Detection System) on the same board using the Python and OpenCV libraries in an attempt to prevent severe traffic accidents. A mathematical model for image processing is implemented by the IIVSS, and it includes operators that alter images based on the characterisation. IIVSS uses various security components to bring the best outcome of the proposed system. We implement the data enhancement technique to increase the number of datasets including pictures of human faces to develop this system. The proposed system is examined on standard datasets and self-generated datasets. The outcome of the experiment of IIVSS shows the accuracy is 95.5% on standard datasets such as LFW, CelebA, and Multi-PIE, etc., whereas 97.4% on the self-generated dataset. It can be observed that the proposed Vehicle security system based on the IIVSS model gives favourable outcomes in aspects of consistency, robustness, and fault tolerance. IIVSS model performance was evaluated through Confusion matrix, ROC curve (Receiver Operating Characteristics Curve), Graph generation for Training loss vs. epoch graph using the different optimizers, and Generating graphical representation of Validation loss vs. epoch using the different optimizers which proved model performs exceptionally well with current data set model.