BUILDING A SMART HUMAN RESOURCE MANAGEMENT SYSTEM – FACE RECOGNITION MODULE

Abstract

Face recognition is one of the most challenging research problems in surveillance systems, while it has been applied to a large number of applications. The face recognition pipeline consists of four phases: face detection, face alignment, feature extraction, and classification, where the major is the feature extraction phase. In this thesis, the research is focused on building the face recognition module, which is a part of the "Building a smart human resource management system" project. Since there are a lot of different approaches to building a face recognition system. This research uses the FaceNet pre-train model for feature extraction when the facial image is converted into 128 dimensional embedded vectors. The research also uses the SVM for classification. In addition, the system achieved high accuracy, which is more than 90%.