Unauthorized Access Detection Using Real-Time Face Recognition

Protecting computer systems from unapproved en- try is a paramount necessity in today’s digital landscape. The proposed work describes a live, integrated security solution for detecting unverified access, utilizing webcam imaging, facial analysis, and automated email notifications to significantly boost login protection. The system automatically activates the webcam when a login attempt is made and captures the image of the person at the terminal. OpenCV is employed to analyze the acquired image and confirm the existence of a human face within the frame. Simultaneously, the user’s geographical location is obtained using an IP-based API. The system then validates the entered username and password with predefined credentials. Should repeated password failure occur, or if the face captured by the camera differs from the authorized user’s profile, the system flags the activity as suspicious or rogue and dispatches an immediate email warning to the system owner with the captured image, time, and location details. In case the real owner forgets the password, a password reset email is also triggered. This multi-faceted defense methodology merges computer vision capabilities with network communication for immediate monitoring and preemptive alerts concerning security breaches, thereby strengthening system access security.