Please use this identifier to cite or link to this item:
Title: Facial feature comparison between biological relatives
Authors: Teoh, Geraldine Shi En
Keywords: DRNTU::Engineering
Issue Date: 2016
Abstract: Facial feature recognition techniques that allow fast and efficient face detection have gained popularity in recent years. Based on the concepts of Viola-Jones robust real-time face detection, this project aims to find facial features that are able to determine if two individuals are related. In this project, images sets of related and unrelated individuals was collected. Facial features such as the nose and eyes were extracted from the image and used for analysis. Other facial features such as the ratios between facial landmarks were also used for analysis. Methods of comparison were used to determine the effectiveness of the facial features in differentiating related and unrelated individuals. In order to differentiate related and unrelated pairs, a threshold value had to be determined for each facial feature. Since not all facial features are equally effective, a weightage was allocated to each facial feature according to their effectiveness in differentiating related and unrelated individuals. The experimental results showed that using a combination of certain facial features with the appropriate threshold values and weightages, related and unrelated individuals were able to be differentiated most of the time. However, more extensive research should be done on a larger sample size to give more substantial results.
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
Facial Feature Comparison Between Biological Relatives.pdf
  Restricted Access
2.09 MBAdobe PDFView/Open

Page view(s)

Updated on Jun 24, 2021


Updated on Jun 24, 2021

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.