Smart Door Keeper

With the advancements in computer vision, computers are now able to recognize and process various objects present in images. This technology has the potential to make life more convenient and improve society. Facial recognition is a task that computers perform better than humans, especially when searching through large databases of faces. In this paper, we present a system that uses computer vision to enhance home security. When someone presses the doorbell, the system detects the person in front of the door. If the person is recognized by the system, then the door can be unlocked. However, if the person is not known, the system will search for logos to identify the delivery company the person represents. If there are no logos detected, then the person is marked as an unknown guest, and the homeowner is notified. Our system includes a front-end, which allows the homeowner to add new people to the database, so on their next visit, the system can recognize them and send a notification accordingly. Additionally, we implemented algorithms for face recognition based on Histogram of Oriented Gradients (HOG) and object detection for identifying the lo-gos of different companies using Oriented FAST and Rotated BRIEF (ORB). This paper presents our system design, software algorithms, implementation, and results, along with a review of relevant work in the field of home automation systems. The traditional system only achieves alerting the home by the sound of a doorbell but nothing regarding identification, record keeping, and surveillance. The agenda is to create a smart doorbell system that incorporates the aspects of record keeping, home automation, and security.