Despite the limitations of current technology, facial recognition from still or moving sources is developing into a useful tool for law enforcement, security, and counterterrorism applications. Currently, facial recognition shows promise and has been used in a small number of applications, The effectiveness of a facial recognition algorithm is examined in this assessment, specifically with regards to factors such as age, gender, ethnicity, facial characteristics, and lighting levels. Face recognition is being used more frequently as a non-intrusive identity verification technique. In this paper, we propose for the first time a strongly privacy-enhanced face recognition system that effectively uses secure multiparty computation techniques to conceal both the biometrics and the result from the server performing the matching operation. These issues can be easily solved by Explainable AI (XAI), and the solutions are clear to the end users. The proposed system is entirely implemented on a Raspberry Pi, enabling a full embedded application. With the aid of open-source libraries for training, defining, and running machine learning models like Tensorflow.js, Keras, and OpenCV, the application was created using Python and HTML on PyCharm/Visual Studio Code. An entire application can be run on a microcontroller, like the Raspberry Pi, which enables plug-and-play operation of the system at any time. The face detection subsystem, which is based on an improved PCA algorithm, can identify faces in moving vehicles and communicate with the vehicle's owner via GSM. The car's location is sent via GPS. This is the most reliable car security system because we can still locate the car even when it is lost and we can see the thief's face. The owner of the car is free to give the password to new drivers as well. This makes the car security system safer and more comfortable.