Face Recognition with Principal Component Analysis (PCA) Application Using Euclidean Distance Measurement

Abstract

Abstract— Face recognition algorithms can be categorized into
5 based methods of linear and non-linear projection, namely:
artificial neural network-based method of non-linear, Gabor
filters and wavelets based methods, fractal-based methods and
methods based on thermal and hyperspectral. PCA is a statistical
method that can explain the formulation of artificial neural
networks and is designed to process the multidimensional
information. With the PCA method to do efficient calculation
where multidimensional information can be simplified into a
number of variables, dimensions and factors serve as the basic
component. Many researchers use PCA method that allows
modeling of a human face by using the parameters in limited
quantities. One advantage PCA method is the ability to process
high-dimensional data modeling that cannot be done by many other
methods because it requires a covariant matrix inverse. PCA is a
better method than matching pursuit (MP), especially on the use of
time, fast and efficient. The purpose of this study was to analyze the
images using image recognition algorithms to calculate the distance
euclidean PCA. The result of this research is the image
recognition can be performed using PCA algorithm to form a
basis vector as the basis for calculating the normalization of an
image. Euclidean distance calculations will provide clarity
regarding the degree of similarity and dissimilarity drawing a
picture