Smart Anti-Theft Door locking System

Privacy and security are two pivotal rights in day-to-day life. At present, keys, passwords and PIN's are used to secure the confidential data. However the above mentioned methods can be compromised and thus propose threats to security. This paper provides an advanced method to enhance the security system using face detection and recognition algorithms integrated with raspberry pi that is used to control the access to the door. Since face is indubitably related to an individual, it cannot be duplicated. This paper consists of three subsystems-Face detection, Feature extraction and Face recognition for door access. Initially the system is trained with authorized persons features, stored in the database. Firstly, the process is started by capturing the image of an object using raspberry pi camera followed by face detection done using Viola Jones algorithm as it provides a greater accuracy in real-time object detection. Next the feature extraction and face detection is done using Local Binary Pattern (LBP) algorithm that can extract local neighboring texture information of grey scale image and can efficiently differentiate between object and background. The extracted features are dimensionally reduced using Principal Component Analysis (PCA) algorithm .The detected face is compared against the stored features and if there is a match the access is provided to the authorized person. If not, the access to the door is denied and an alarm is raised alerting the admin.