Door Access Control based on Illumination Invariant Face Recognition in Embedded System

The use of conventional keylock is sometimes insecure due to it can still be hijacked by a thief. In this paper, we address the problem by offering a biometric-based keylock system based on automatic face recognition. We proposed an innovation into our face recognition based electronic keylock system by an ability to recognizing person even in a harsh illumination condition. An illumination invariant face recognition is adopted to robustly recognize a person even there is a change in illumination. This method is simple, and it can be embedded into a low-cost embedded system, like Raspberry Pi or Lattepanda. Based on our experimental results on Lattepanda, our system shows good performance and is very promising to improve the ability of conventional electronic keylock systems.