Genetic algorithm (GA) can improve the classification of the face recognition process in the principal component analysis (PCA). However, the accuracy of this algorithm for the smart home security system has not been further analyzed. This paper presents the accuracy of face recognition using PCA-GA for the smart home security system on Raspberry Pi. PCA was used as the face recognition algorithm, while GA to improve the classification performance of face image search. The PCA-GA algorithm was implemented on the Raspberry Pi. If an authorized person accesses the door of the house, the relay circuit will unlock the door. The accuracy of the system was compared to other face recognition algorithms, namely LBPH-GA and PCA. The results show that PCA-GA face recognition has an accuracy of 90 %, while PCA and LBPH-GA have 80 % and 90 %, respectively.

As industrial 4.0 advances such as Internet of Things (IoT), Big Data, Cloud Computing , and Artificial Intelligent  (AI) has prompted reseachers to innovate in various fields, including security systems. The security system is an important issue due to rise of theft in a residence. A security system is needed for home authorizon to prevent the crime of theft. Security systems are built using facial recognition. The research proposes to develop a security system using facial recognition  based Raspberry Pi with Python programming and utilize the OpenCV. System testing includes training function testing, facial recognition function, image delevery function, decision-making function, and system performance testing, system performance for facial recognition is calculated using confusion matrix formula that produces 100% sensitivity, 13% specitificity, and 97% accuracy.