Graduate Project
 

Enhancing Facial Recognition: A Comparative Analysis of Image Preprocessing Techniques in PCA, KL Expansion and SVD

Público Deposited

Conteúdo disponível para baixar

Baixar PDF
https://ir.library.oregonstate.edu/concern/graduate_projects/3197xw046

Descriptions

Attribute NameValues
Creator
Abstract
  • Facial recognition has become increasingly important in recent years, due to the wide range of applications it has in fields such as security, surveillance, and human-computer interaction. Three popular methods for facial recognition are the Principal Component Analysis (PCA), Karhunen-Loeve Expansions, which is fundamentally a continuous form of PCA but with stochastic random processes, and Singular Value Decomposition. This study explores the use of interpolation and resampling techniques on facial recognition accuracy using Principal Component Analysis (PCA) and Singular Value Decomposition (SVD). Images were preprocessed by mean subtraction, followed by linear interpolation, cubic spline interpolation, and resampling to standardize their dimensions. The processed images were then subjected to SVD to extract eigenfaces. The analysis revealed that cubic spline interpolation better preserves facial details critical for recognition compared to resampling. It also revealed that performing a linear interpolation is equivalent to resampling the data. This finding underscores the importance of choosing suitable preprocessing methods in facial recognition systems, with implications for enhancing their accuracy and applicability in various domains.
License
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Declaração de direitos
Publisher
Peer Reviewed
Language

Relações

Parents:

This work has no parents.

Itens

Miniatura Título Data de carga Acesso Ações
file details TuffourPrinceA2024.pdf 2024-01-24 Público Baixar