Open AccessPengenalan Citra Wajah Dengan Menggunakan Transformasi Wavelet Diskrit dan Jaringan Saraf Tiruan Back-Propagation
Author(s) -
Suhendry Effendy
Publication year - 2010
Publication title -
comtech/comtech
Language(s) - English
Resource type - Journals
eISSN - 2476-907X
pISSN - 2087-1244
DOI - 10.21512/comtech.v1i2.2569
Subject(s) - artificial intelligence , computer science , computer vision , pattern recognition (psychology) , face (sociological concept) , noise (video) , artificial neural network , image (mathematics) , wavelet , discrete wavelet transform , facial recognition system , wavelet transform , social science , sociology
This paper discusses the facial image recognition system using Discrete Wavelet Transform and back-propagation artificial neural network. Discrete Wavelet Transform processes the input image to obtain the essential features found on the face image. These features are then classified using an back-propagation artificial neural network for the input image to be identified. Testing the system using facial images in AT & T Database of Faces of 400 images comprising 40 facial images of individuals and web-camera catches as many as 100 images of 10 individuals. The accuracy of level of recognition on AT & T Database of Faces reaches 93.5%, while the accuracy of level of recognition on a web-camera capture images up to 96%. Testing is also done on image of AT & T Database of Faces with given noise. Apparently the noise in the image does not give meaningful effect on the level of recognition accuracy.