Welcome To UTPedia

We would like to introduce you, the new knowledge repository product called UTPedia. The UTP Electronic and Digital Intellectual Asset. It stores digitized version of thesis, dissertation, final year project reports and past year examination questions.

Browse content of UTPedia using Year, Subject, Department and Author and Search for required document using Searching facilities included in UTPedia. UTPedia with full text are accessible for all registered users, whereas only the physical information and metadata can be retrieved by public users. UTPedia collaborating and connecting peoples with university’s intellectual works from anywhere.

Disclaimer - Universiti Teknologi PETRONAS shall not be liable for any loss or damage caused by the usage of any information obtained from this web site.Best viewed using Mozilla Firefox 3 or IE 7 with resolution 1024 x 768.

FACE CLASSIFICATION FOR AUTHENTICATION APPROACH BY USING WAVELET TRANSFORM AND STATISTICAL FEATURES SELECTION

DAWOUD JADALAH, NADIR NOURAIN (2011) FACE CLASSIFICATION FOR AUTHENTICATION APPROACH BY USING WAVELET TRANSFORM AND STATISTICAL FEATURES SELECTION. Masters thesis, UNIVERSITI TEKNOLOGI PETRONAS.

[img]
Preview
PDF
Download (2021Kb) | Preview

Abstract

This thesis consists of three parts: face localization, features selection and classification process. Three methods were proposed to locate the face region in the input image. Two of them based on pattern (template) Matching Approach, and the other based on clustering approach. Five datasets of faces namely: YALE database, MIT-CBCL database, Indian database, BioID database and Caltech database were used to evaluate the proposed methods. For the first method, the template image is prepared previously by using a set of faces. Later, the input image is enhanced by applying n-means kernel to decrease the image noise. Then Normalized Correlation (NC) is used to measure the correlation coefficients between the template image and the input image regions. For the second method, instead of using n-means kernel, an optimized metrics are used to measure the difference between the template image and the input image regions. In the last method, the Modified K-Means Algorithm was used to remove the non-face regions in the input image. The above-mentioned three methods showed accuracy of localization between 98% and 100% comparing with the existed methods. In the second part of the thesis, Discrete Wavelet Transform (DWT) utilized to transform the input image into number of wavelet coefficients. Then, the coefficients of weak statistical energy less than certain threshold were removed, and resulted in decreasing the primary wavelet coefficients number up to 98% out of the total coefficients. Later, only 40% statistical features were extracted from the hight energy features by using the variance modified metric. During the experimental (ORL) Dataset was used to test the proposed statistical method. Finally, Cluster-K-Nearest Neighbor (C-K-NN) was proposed to classify the input face based on the training faces images. The results showed a significant improvement of 99.39% in the ORL dataset and 100% in the Face94 dataset classification accuracy. Moreover, a new metrics were introduced to quantify the exactness of classification and some errors of the classification can be corrected. All the above experiments were implemented in MATLAB environment.

Item Type: Thesis (Masters)
Subject: UNSPECIFIED
Divisions: UNSPECIFIED
Depositing User: Users 5 not found.
Date Deposited: 05 Jun 2012 11:28
Last Modified: 25 Jan 2017 09:42
URI: http://utpedia.utp.edu.my/id/eprint/3053

Actions (login required)

View Item View Item

Document Downloads

More statistics for this item...