Implementasi Pengenalan Wajah menggunakan Ekstraksi Fitur Patterns Of Oriented Edge Magnitudes

Nadya, R.AY. Noormala (2019) Implementasi Pengenalan Wajah menggunakan Ekstraksi Fitur Patterns Of Oriented Edge Magnitudes. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

Text 05111440000127-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2022. Download (1MB) | Request a copy

Abstract

Demi meningkatkan keamanan di daerah umum, pengenalan wajah menjadi salah satu yang dipertimbangkan. Pengenalan wajah merupakan klasifikasi pola yang sulit karena adanya inter-cluster variation (variasi citra wajah pada orang yang sama). Variasi pencahayaan merupakan salah satu faktor utama dalam inter-cluster variation. Variasi pencahayaan membuat perubahan drastis dalam tampilan sebuah wajah.
Dalam Tugas Akhir ini diimplementasikan pengenalan wajah menggunakan metode ekstraksi fitur Patterns of Oriented Edge Magnitudes (POEM). Metode klasifikasi yang dipakai adalah Support Vector Machine (SVM). Sebelum dilakukan ekstraksi fitur, terlebih dahulu dilakukan frontalization dan Difference of Gaussians (DoG) untuk memperbaiki kualitas citra.
Uji coba yang dilakukan menggunakan dataset FERET. Hasil uji coba yang menggunakan metode ekstraksi fitur POEM dengan klasifikasi SVM mendapatkan akurasi terbaik pada kelas Fb sebesar 79.1%, Fc sebesar 89.1%, Dup1 68.8%, dan Dup2 61.8%.
================================================================================================
In order to improve security in the general area, face recognition is one of the ones that supports them. Variations between clusters (variations in facial image on the same person). Lighting variation is one of the main factors in variations between clusters. Variations make drastic changes in facial appearance.
In this Final Project, will be applied face recognition using the feature extraction method Pattern Oriented Edge Magnitudes (POEM). The classification method used is Machine Vector Support (SVM). Before feature extraction, frontalization and Difference of Gaussians (DoG) were carried out to improve image quality.
Trials carried out using the FERET dataset. The results of the trial using the POEM feature extraction method with SVM classification obtained the best accuracy of Fb class gives 79.1%, Fc gives 89.1%, Dup1 68.8%, and Dup2 61.8%.

Item Type:	Thesis (Undergraduate)
Additional Information:	RSIf 006.42 Nad i-1 2019
Uncontrolled Keywords:	rontalization, Patterns of Oriented Edge Magnitudes, Difference of Gaussians, Support Vector Machine
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA1650 Face recognition. Optical pattern recognition.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	R.AY. NOORMALA NADYA
Date Deposited:	25 Jan 2022 01:56
Last Modified:	25 Jan 2022 01:56
URI:	http://repository.its.ac.id/id/eprint/61675

Actions (login required)

View Item