Review Paper: Face Recognition Techniques in Digital Image Processing

Abstract

Purpose: This paper aims to explore the face recognition process, the accuracy of the face recognition system, and the system can recognize face in real time. Background: Face recognition is a biometric technique that allows a computer to recognize faces by comparing the input image with the image from the provided database. Currently, face recognition has been widely used in various fields and one of these usages is for security systems. In a security system, a high accuracy and very good performance in the face recognition process is needed. Design/Methodology/Approach: . There are many methods used in face recognition systems, some of them are Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Nearest Neighbor, Gabor Wavelet and Boosted Cascade Classifier. There are several factors that influence the success of the face recognition system, namely the distance of the face to the device such as camera, lighting, the number of images of people's faces stored and the performance of the computer used. Results/Findings: The accuracy and performance values from the PCA, LDA, K-Nearest Neighbor, Gabor Wavelet, Boosted Cascade Classifier method with several test samples according to the factors that influence success in the face recognition system will be exposed. Conclusion and Implications: This paper is expected to make readers who want to build systems with face recognition features easier to determine which method will produce the appropriate level of accuracy without reducing the performance of the system built.