We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
DIGITAL IMAGE PROCESSING APPROACH TO FINGERPRINT AUTHENTICATION.
- Authors
RAKUN, J.; BERK, P.; STAJNKO, D.; OCEPEK, M.; LAKOTA, M.
- Abstract
In this work, we present a digital image processing approach to fingerprint based authentication, which could be a part of an embedded robotic system. It uses digitalized fingerprint images that are made up by unique distribution of minutiae features such as ridge endings, ridge bifurcations, spurs, delta and cores. In order to minimize the computational requirements for an embedded system, the approach transforms each digitalized image to a distinct 12 dimensional 8-bit feature vector, which is accomplished with the help of gradients and wavelet transform. The feature vectors are then simply compared and used to compute a MSE that serves as a similarity measure. We have analysed 30 randomly selected examples and results show the approach is suitable to be used as a part of an embedded system. By using a small test database we have reached a success rate of 98.7 %.
- Subjects
DIGITAL image processing; HUMAN fingerprints; EMBEDDED computer systems; COMPUTER access control; IMAGE analysis; DATABASES
- Publication
DAAAM International Scientific Book, 2012, p517
- ISSN
1726-9687
- Publication type
Article
- DOI
10.2507/daaam.scibook.2012.43