A Near Optimal Smart Door Security System Based on Internet of Things

This project aims to solve the security problem faced by digital door locking system. By integrating technologies such as computer vision, image processing, internet of things and deep learning various solutions can be developed for a digital to locking system. Most of the existing system fails to recognize the person who wears a mask aur covered in mud etc. This was due to minimum accuracy and reliability. In this regard the proposed approach uses deep learning method to recognize a face using mask or mud. The proposed approach also comes with a cost effective iot based system using principal component analysis. Modifications have been done in Eigen face algorithm which is used by proposed approach. This algorithm is used in the case of component images which are blurred and get clear face recognition automatically. During the comparison it has been identified that proposed approach performs well then the existing approaches.