A Computer Vision-Based Face Lock for Automotive Ignition

Vehicle security remains a paramount concern, with traditional key-based systems being susceptible to theft and unauthorized access. This paper presents the design and implementation of a prototype for a secure, biometric-based car ignition system using facial recognition. The system leverages a computationally efficient computer vision algorithm, Local Binary Patterns Histograms (LBPH), to perform face recognition. The core process involves capturing a driver’s face via a webcam, authenticating it against a pre-enrolled database of authorized users, and subsequently sending a control signal to an embedded subsystem. The ignition control hardware, built around an ESP32 microcontroller, uses a relay to switch a DC motor, simulating the vehicle’s engine starter. Communication between the host computer performing the recognition and the ESP32 is established via a standard USB port. This framework demonstrates a low-cost, practical, and enhanced security alternative to conventional car ignition systems.