A Multi-Modal Approach to Frontal and Non Frontal Facial Feature Detection

Master's Thesis

Abstract

Biometrics are measurable characteristics specific to an individual. Theycan be used to identify individuals. The use of biometrics for identification hasthe potential to make our lives easier, and the world we live in a safer place.While numerous different types of exploitable biometrics exist, facialidentification is highly pursued because the data can be captured at a distancewithout requiring the subject s active cooperation. While traditionally 2D imagesof faces have been used, 3D scans that contain both 3D data and registered colorare becoming easier to acquire.Before 3D face images can be used to identify an individual, they requiresome form of initial alignment information, typically based on facial featurelocations. This thesis proposes and analyzes a multimodal approach to automaticfacial feature detection. After beginning with a discussion on biometrics and therole of automatic facial feature detection, we provide a comparative evaluation of3D sensors. We follow this by a discussion of the algorithm s performance whenconstrained to frontal images and an analysis of its performance on a morecomplex dataset with significant head pose variation.

Attributes

Attribute NameValues
URN
  • etd-12142005-093514

Author Christopher Bensing Boehnen
Advisor Patrick Flynn
Contributor Kevin Bowyer, Committee Member
Contributor Patrick Flynn, Committee Chair
Contributor Robert Stevenson, Committee Member
Degree Level Master's Thesis
Degree Discipline Computer Science and Engineering
Degree Name Master of Science in Computer Science and Engineering
Defense Date
  • 2005-11-21

Submission Date 2005-12-14
Country
  • United States of America

Subject
  • 3D Imaging

  • 3D Face

  • feature

  • Accuracy

  • 3D Scanning

  • biometrics

  • multi-modial

Publisher
  • University of Notre Dame

Language
  • English

Record Visibility Public
Content License
  • All rights reserved

Departments and Units

Digital Object Identifier

doi:10.7274/h128nc6074f

This DOI is the best way to cite this master's thesis.

Files

Please Note: You may encounter a delay before a download begins. Large or infrequently accessed files can take several minutes to retrieve from our archival storage system.