PENGENALAN WAJAH MANUSIA MENGGUNAKAN ANALISIS KOMPONEN UTAMA (PCA)DAN JARINGAN SYARAF TIRUAN PERAMBATAN-BALIK

Zayuman , Hidayat and Santoso , Imam and Isnanto, R.Rizal (2011) PENGENALAN WAJAH MANUSIA MENGGUNAKAN ANALISIS KOMPONEN UTAMA (PCA)DAN JARINGAN SYARAF TIRUAN PERAMBATAN-BALIK. Undergraduate thesis, Jurusan Teknik Elektro Fakultas Teknik Undip.

[img]
Preview
PDF - Published Version
146Kb

Abstract

Face is a part of human body that has unique characteristics, so that they can differentiate and recognize someone just by look their face. Up till now the research on biometrical system who was carried out of the university student is not completely delve yet of human characteristics with automatic face recognized on computer system. For those reasons, a research is required to be done to make a face recognition system, with the result of the system, in this case a computer can recognize or identify someone with face image like a human. The steps of application system are face image inputing, feature extraction and reduction used Principal Component Analysis (PCA), training used artificial feedforward and backpropagation neural network. The processing test was obtained using training to data test, and then the result, of output value and based data target value was compared. Data with appropriate value between data training class and input image class selected as a recognized face. The research shows that combination of PCA and backpropagation Artificial Neural Network is good enough for facerecognition system. That is showed on introducing level during testing, with introducing level average is 85 %. The principal component can lessen up to 7 principal component. With the result is equal to the 60 principal component on the testing with ORL data and four training images. Best successful level of the system is 96,67% using input device 8 mega pixels Olympus digital camera. The while research of Wibowo[19] intake data used digital camera on cellular phone Nokia 6600 and by downloading from internet, and grade of recognition of testing is 95 %. The principal component can lessen up to 30 principal component. With the result is equal to the 60 principal component. Keywords - Face recognition, Principal Component Analysis (PCA), Artificial backpropagation neural network, training images, testing images.

Item Type:Thesis (Undergraduate)
Subjects:T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions:Faculty of Engineering > Department of Electrical Engineering
Faculty of Engineering > Department of Electrical Engineering
ID Code:25291
Deposited By:Mr. Sudjadi Pranoto
Deposited On:10 Jan 2011 18:43
Last Modified:10 Jan 2011 18:43

Repository Staff Only: item control page