Perbandingan pengenalan wajah berbasis fitur dan berbasis citra dengan praproses analisis komponen utama
Abstract
Face recognition is one of many important researches and many applications have implemented it. Two different techniques have been proposed for face recognition. There are feature-based systems and image-based systems. The objective of this research is to compare two simple strategies between the systems. The first system is based on a set of geometrical features such as nose width and length, mouth position, and the second one is based on almost grey-level pixel based. The data are faces of 20 people: 3 females and 17 males with 6 images each. Face data are usually difficult to recognize, classify and analysis. This research uses as Principal Component Analysis (PCA) method to reduce face’s data. K-Nearest Neighbor algorithm is used to examine the face data’s accurate and validation from both methods.The results obtained on testing sets is about 60% correct using geometrical features and perfect recognition using image-base system is about 95%. Noise is also used in this research with several varians such as 0,02 , 0,10 , and 0,20. It is difficult to recognise image with high varians of Salt and Pepper noise in image-based systems.