Open Access||Research ArticleSecurity System Using Face Recognition: Machine Learning Based Approach
Department of Electronics and Telecommunications Engineering, Sinhgad Institute of Technology, Lonavala, 410401, Maharashtra, India
Abstract
The rising occurrences of illegal entry and security breaches have made guaranteeing safety inside residential societies a key worry in the contemporary day. Often lacking in consistent and tamper-proof access control are traditional security solutions like manual guarding, RFID cards, or keypad locks. These traditional approaches are vulnerable to human error, duplication, and illegal use, hence stressing the critical need for a more smart and automated solution. Recent studies in the area of machine learning and computer vision have produced encouraging findings in facial recognition technologies. Research suggests building a Society Security System Using Face Recognition Technique and Machine Learning to solve these issues, with the goal of producing a low-cost, efficient, and frictionless access control system. Using a facial recognition model coupled with Support Vector Machine (SVM) classification, the proposed security system accurately identifies authorised faces. Testing on a dataset of authorised and unauthorized individuals revealed an overall accuracy of 95%. The model showed good real-time performance, fast response time, and resilience to changing lighting conditions and facial emotions. Attempts at unauthorized access were efficiently spotted and denied, hence guaranteeing improved security.
Novelty Statement
Using a facial recognition model coupled with Support Vector Machine (SVM) classification, the proposed security system accurately identifies authorised faces.