Auto Safety Technology With Enhanced Facial Recognition To Prevent Replay Attacks

One of the most popular uses of IoT technology today is in the automotive industry, where automobiles can be made intelligent. Yet regrettably, there are several crimes involving these automobiles. So, it has become extremely difficult for the IoT to prevent such crimes from skilled burglars. This paper proposes an enhanced facial recognition-based authentication for vehicles. Face recognition technology may be used to allow only authorized users to access the car and send the owner a photo of any unauthorized users. Currently, a variety of machine learning methods are utilized to categorize genuine and fake users. The experimentations are carried out on the standard replay attack dataset. The five assorted machine learning classifiers such as linear regression, multilayer perceptron, simple linear regression, SMOReg, Random Tree, etc. been used in experimentation. TSBTC and the Otsu thresholding algorithm are employed to extract features from the input images. OTSU and TSBTC were merged at the feature level to produce better accuracy, and the results clearly indicate that the performance of the fusion of OTSU and TSBTC is better than individuals.