Biometric and IoT-based Vehicle Security Systems: Advancements and Applications

Biometric authentication is a safe and effective access control method used in today's vehicles. This work discusses an authentication system based on dual-modal verification with face and fingerprint recognition for improved security and convenience to the user. Deep learning is used to train recognition models with a publicly accessible biometric dataset. Preprocessing operations like noise removal, data augmentation, and class balancing are used to make the models robust. Experimental outcomes show that face recognition is found to have more accurate performance than fingerprint recognition, along with a discussion of factors contributing to this performance. To check real-world feasibility, we also test the system's real-time operation, such as time taken for processing and network latency in an IoT-based setup. Security issues, like spoofing attacks, are handled by liveness detection techniques. Comparative study with other biometric modalities, such as iris recognition, shows the efficacy of the presented system.