<![CDATA[Face is a biometric feature that can be identified by most people in the world. Digital images of faces can be utilized by computer systems for the purpose of recognizing one's identity. In doing facial recognition, computers need the right method to get high accuracy with good computation time. In this study, the Linear Discriminant Analysis (LDA) feature extraction method and the K-Nearest Neighbor (KNN) classification method were implemented using the Particle Swarm Optimization (PSO) method for feature selection. The dataset used in this study is the ORL database of 400 face images taken from 40 different people and each person has 10 face images. Based on the test results, the highest accuracy obtained by LDA and KNN is 70.00% with a computation time of 0.2233s, while LDA and KNN based on PSO get the highest accuracy of 71.67% with a computation time of 0.1224s. So, it can be concluded that there is an increase in accuracy after the application of PSO, which is 1.67% and saves computation time by 0.1009s. An increase in accuracy indicates that PSO is able to select the best features correctly and is feasible to be processed at the time of classification.]]>
Copyrights © 2022
