WoodardDL122004.pdf (988.64 kB)
Exploiting Finger Surface as a Biometric Identifier
thesis
posted on 2004-12-14, 00:00 authored by Damon Lamar WoodardBiometrics, the discipline of establishing an individual's identity based upon physical or behavioral characteristics, has become a majorresearch area due to the numerous applications for reliable personal identification. The performance of a biometric system is highly dependent on the chosen biometric identifier. A novel approach for personal identification which utilizes 3D finger surface features as a biometric identifier is presented. Using 3D range images of the hand, a surface representation for the index, middle, and ring finger is calculated andused for comparison to determine subject similarity. The curvature based shape index is used to represent the finger surfaces. A large unique database of hand images supports the research. Data sets obtained over time are used to examine the performance of each individual finger surface as a biometric identifier as well as the recognition performance obtained when combining them. The probe and gallery sets sizes are varied to determine their affect on overall system performance. Performance results for both authentication and identification tasks arepresented, which suggest that 3D finger surface is a viable choice as a biometric identifier.
History
Date Created
2004-12-14Date Modified
2018-10-29Defense Date
2004-09-14Research Director(s)
Patrick J. FlynnCommittee Members
Kevin W. Bowyer Robert L. Stevenson Gregory MadeyDegree
- Doctor of Philosophy
Degree Level
- Doctoral Dissertation
Language
- English
Alternate Identifier
etd-12142004-175627Publisher
University of Notre DameProgram Name
- Computer Science and Engineering
Usage metrics
Categories
No categories selectedKeywords
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC