Face recognition with accessories using Eigenfaces

This study uses a dataset of unoccluded faces for training to examine the effectiveness of using an eigenface approach to identify faces covered by accessories, particularly sunglasses. The robustness of the eigenface method against partial face occlusion is assessed using principal component analys...

Full description

Saved in:
Bibliographic Details
Main Author: Toh, Safarudin Wilyam
Other Authors: Anamitra Makur
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2024
Subjects:
Online Access:https://hdl.handle.net/10356/176949
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:This study uses a dataset of unoccluded faces for training to examine the effectiveness of using an eigenface approach to identify faces covered by accessories, particularly sunglasses. The robustness of the eigenface method against partial face occlusion is assessed using principal component analysis, a method often employed in the field of computer vision to perform successful human face detection. This project intends to show that the suggested face recognition system can correctly identify a person regardless of the presence of accessories, for people who are part of the training set. In contrast, it should be able to classify faces that are absent from the training set as unauthorized with reasonable accuracy.