Vehicle Anti-Theft Systems Using Vision Transformer and Iris Identification

Traditional and existing techniques of car theft protection are insufficient. In this research, we present a technique for verifying driver identity by examining the iris of the eye, an identifier that has been proved to be a unique key identifying individuals in order to improve security and avoid theft. The proposed method starts with obtaining an image of the eye with a camera, followed by a succession of image processing processes. These techniques enable the extraction of the necessary elements for input into a model based on Visual Transformer (ViT), which has recently emerged as a competitor to convolutional neural networks (CNN) in picture categorization. Our method employs a camera linked to a Raspberry Pi, which is linked to the car engine. The engine will only start if the iris of the automobile owner is recognized. This strategy was effective due to the low cost of the suggested model, as well as the response time and accuracy aided by attention mechanisms that combine recent ViT architecture with hyperparameter fine-tuning.