Home Security System Based on Facial Recognition

Security systems serve the major functions of protection and safety. Home security is a major predicament in today's world. It is proven that facial recognition, which is a relatively new technology, is the most efficient way of providing security. The high-end security systems used today are very expensive and use biometrics like fingerprint or iris scanners. These systems require additional equipment for recognition which can be replaced by a simple camera while using facial recognition. Our project is aimed at creating an efficient security system incorporating the concepts of Artificial Intelligence (AI) and Machine Learning (ML) to implement facial recognition. Histograms of Oriented Gradients is the algorithm used for facial recognition. The usage of doorbells is also uprooted to attain a more automated approach to detecting the presence of someone at the door by using facial detection. The above goals are carried out with the use of OpenCV, Imutils and Dlib libraries in Python, with a Raspberry Pi as the processing unit of the security system. The door lock, controlled by a stepper motor, operates immediately once the person detected is recognized. Application of the Internet of Things (IoT) and the Global System for Mobile Communications (GSM) is also incorporated in the situation when the detected person is not recognized.