Perbandingan Akurasi Algoritma Principal Component Analysis Dengan Algoritma Convolutional Neural Network Dalam Pengenalan Wajah

Abstract

Facial recognition technology is used in various fields such as criminal identification, security purposes, finding missing people, diagnosing diseases, forensic investigations, identifying people on social media platforms, opening mobile phones, access control of meeting rooms, bank vaults. This paper presents a performance comparison between PCA and CNN algorithms. The aim of this research is to test the classification and compare the accuracy of PCA and CNN algorithms in face recognition. The method is to test the classification of Euclidean distance weights and compare performance tests which include; precision, recall and accuracy. The results showed that the accuracy of the PCA algorithm predicted TP by 100%, while the CNN algorithm predicted TP by 82%, while FN predicted 0. The recall performance on the PCA algorithm predicted TP by 100%, while on the CNN algorithm the recall performance predicted TP by 82%.  Accuracy in the PCA algorithm is known that the TP prediction result is 100%, where TN, FN and FP are 0, while the accuracy performance in the CNN algorithm, TP is 82%, FN is 18%, TN and FP are 0.