Copyright © 2018 Authors. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. International Journal of Engineering & Technology, 7 (4.10) (2018) 55-58 International Journal of Engineering & Technology Website: www.sciencepubco.com/index.php/IJET Research paper Implementation of Facial Recognition for Home Security Systems Arnab Pushilal 1 , Sulakshana Chakraborty 2 , Raunak Singhania 3 , P. Mahalakshmi 4 * 1,2,3, 4, School of Electrical Engineering, Vellore Institute of Technology, Vellore, India, 632 014 *Corresponding author E-mail: pmahalakshmi@vit.ac.in Abstract In this paper, the design and development of a home security system has been detailed which uses facial recognition to conform the iden- tity of the visitor and taking various security measures when an unauthorized personnel tries accessing the door. It demonstrates the im- plementation of one of the most popular algorithm for face recognition i.e. principal component analysis for the purpose of security door access. Since PCA converts the images into a lower dimension without losing on the important features, a huge set of training data can be taken. If the face is recognized as known then the door will open otherwise it will be categorized as unknown and the microcontroller (Arduino Uno) will command the buzzer to start ringing. Keywords: Arduino Uno; Covariance, Eigenvalue; Eigenvector; Eigen face; Eigen vector; Euclidean distance; Manhattan distance; PCA 1. Introduction With the advancement in the areas of personal identification in access control and technologies such as biometric identification seem to be gaining more popularity over the use of cards, pattern or password. Such system also requires the individual to touch the hardware for identification which is not the case in face recogni- tion systems and hence has an added advantage making it a much quicker and efficient process [1]. In this system facial recognition has being implemented by the means of Principal component analysis. This approach is preferred because of its simplicity, speed and learning capacity. In PCA this the image are reduced to a lower dimensional representation consisting of only the salient features [2]. These salient features are called Eigen faces because these are the eigen vectors or the principal components of the training set images. These do not necessarily correspond to nose, eyes or other such part of the face. The images are categorized by finding out the weighted sum of eigen faces so for recognizing a face all we have to do is compare these weights with the faces of the known individuals [3]. In this paper we have proposed a system which comprises of di- mensional reduction of the images, facial recognition, classifica- tion and a door access security system [4]. To access control to the door, an image will be captured and will be compared with the stored images using PCA after which the distance of the captured image from the stored images will be calculated to decide whether the person is authorized to gain access or not [5]. In this paper we have used two methods for calculating such distances namely Manhattan and Euclidian distance. Comparison of the two dis- tances has been done which has been further discussed in the pa- per [6]. When the person has been classified as known the servo motor would rotate a certain angle signifying access to the door where as if the person is categorized as unknown, the system takes security measures such as raining the alarm with the blinking of the red led [7]. In this system, PCA and face classification has been implemented with the help of MATLAB installed on PC. An Arduino Uno board has been used which is interfaced with the PC via the USB port. The rest of the hardware connections are shown below. Fig.1. Block diagram of the facial recognition system for home security systems 2. Methodology 2.1. Principal Component Analysis The Principal component analysis (PCA) is a statistical approach that performs orthogonal transformation on a set of possibly corre- lated variables and converts them into a smaller set of linearly uncorrelated variables called the principal components which is done by extracting the most important features of the data set and