Metode Ekstraksi Ciri untuk Membedakan Citra Wajah Asli dan Foto Berbasis Perceptron

Abstract

Face is one of media for human identification. Previous studies aimed at identifying human face were for a two-dimensional images. Thus, fraud may occur when providing input in two-dimensional face images (photos). This study aims to distinguish the original three-dimensional face image with two-dimensional face image. Feature extraction based on facial geometry principles (Incomplete sentence, subject only, do not know what the authors mean). Face images (both the original and the photos) were captured at deviated angle, to the left and to the right. Each image is then sliced for each face components (eyes and nose) and sought the position of the center point of each component. Comparison between the value of the right eye-nose projection vector to the left-right eye vector and the value of the left-right eye vector become the characteristics of each image. The perceptron method was used for the classifiers. The result, the software can distinguish the original three-dimensional and two-dimensional face image with an error of 8.33% of the 24 tested images. Error occurred for some samples that show big round nose.