Abstract
In this paper, we present a statistical approach to core detection in fingerprint images that is based on the likelihood ratio, using models of variation of core templates and randomly chosen templates. Additionally, we propose an alternative dimension reduction method. Unlike standard linear discriminant analysis (LDA), this method is able to account for differences in both the mean and the covariance matrix of two classes. We show that the method is able to correctly detect the core position in 95% of the fingerprints.
Original language | Undefined |
---|---|
Title of host publication | 4th IEEE Benelux Signal Processing Symposium (SPS-2004) |
Place of Publication | secretariat Delft |
Publisher | IEEE |
Pages | 41-44 |
Number of pages | 4 |
Publication status | Published - Apr 2004 |
Event | 4th IEEE Benelux Signal Processing Symposium, SPS 2004 - Hilvarenbeek, Netherlands Duration: 15 Apr 2004 → 16 Apr 2004 Conference number: 4 |
Conference
Conference | 4th IEEE Benelux Signal Processing Symposium, SPS 2004 |
---|---|
Abbreviated title | SPS |
Country/Territory | Netherlands |
City | Hilvarenbeek |
Period | 15/04/04 → 16/04/04 |
Keywords
- SCS-Safety
- EWI-756