Smart Security System for Theft Protection Using Face Recognition

Abstract

The number of car robbery attempts at the local and international scale is rising rapidly in this modern era. By inventing robbery techniques, the owners are afraid that their cars will be robbed from their ordinary parking lot or from outside. This makes vehicle protection against robbery important as a result of insecurity. The computer vision based real-time vehicle safety system solves this problem. The proposed car safety system carries through real time user authentication based on image processing using face detection and recognition techniques and a microprocessor-based control system attached to the car. The infrarot sensor attached to the driver's vehicles seat activates the hidden camera, which is fixed inside the vehicle, as the person enters the parked vehicle overcoming the existing security features. The person's face is detected using Viola Jones algorithm once the image is obtained from the activated camera. The extracted face is recognized using the improved Linéar Discriminant Analysis (LDA) algorithm that distinguishes many features rather than looking for an exact pattern based on the Euclidean distance. Authorization requires that the threshold value is established and compared to the Euclidean distance over which the person is not authenticated. The face is sent to the mobile of the owner as an MMS via the operating GSM modem, which is classified as unknown. The owner shall be controlled with the relay in accordance with the owner’s command when the information is received. The way to authenticate the person would be efficient and efficient in terms of vehicle safety.