Pengujian Pengenalan Wajah Dengan Menggunakan Algoritma K-Nearest Neighbor

Abstract

 Nowadays, the development of science and technology related to the measurement and statistical analysis of biometric data is growing so rapidly. The face is the first tool to identify a person. However, the facial recognition process is performed by a computer is not as easy as facial recognition done by humans. Facial recognition processes on the computer in addition to slow, also very influential on external factors such as light, the position of the face, as well as accessories being worn. Because there are different factors that affect face recognition, it is used methods such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and Locality Preserving Projections (LPP) to extract the features contained in the image to be recognized more easily. After feature extraction, classification shall be made by the method of k-nearest neighbor (KNN) as a process of getting to know a person's face. The results of some tests performed, LDA has the lowest decision error rate compared to two other methods tested. Where the average error of the LDA was not until seven errors. Feature extraction method is best for optimizing KNN classification method is LDA. LDA method is excellent in extracting features that exist in the image sehinnga facilitate KNN method in classifying the identity of the image being tested.