PAPERS
To return to the "Published Papers" main page, click here.
THE SEVENTH INTERNATIONAL CONFERENCE ON FORENSIC COMPUTER SCIENCE - ICoFCS 2012
Print ISBN 978-85-65069-08-3 - Online ISBN 978-85-65069-06-9, pages 41-45
DOI: 10.5769/C2012007 and http://dx.doi.org/10.5769/C2012007
Fingerprint Image Quality Estimation Using a Fuzzy Inference System
By Carlos Henrique Moniwa Tada, Alexandre Zaghetto, and Bruno Macchiavell
To download this paper, click here.
Print ISBN 978-85-65069-08-3 - Online ISBN 978-85-65069-06-9, pages 41-45
DOI: 10.5769/C2012007 and http://dx.doi.org/10.5769/C2012007
Fingerprint Image Quality Estimation Using a Fuzzy Inference System
By Carlos Henrique Moniwa Tada, Alexandre Zaghetto, and Bruno Macchiavell
To download this paper, click here.
ABSTRACT
Fingerprint image quality affects the performance of automatic fingerprint recognition systems, such as AFIS (Automated Fingerprint Identification System). This work proposes a new method for fingerprint image quality estimation using local features (contrast, curvature and ridge flow). A fuzzy inference system is used to combine these features into a single image quality score. Tests are carried out with fingerprint images from the Fingerprint Verification Competition 2006 (FVC 2006) DB2-A database and fingerprint matching software BOZORTH3 from NIST Biometric Image Software (NBIS). After the removal of 5%, 10% and 15% of the poorest quality fingerprints from the DB2-A, we obtained an improvement of 30.6%, 32.6% and 37.9% in EER (Equal Error Rate), respectively.
KEYWORDS
Fingerprint Image Quality, Fuzzy Inference Systems, Digital Image Processing.
Fingerprint image quality affects the performance of automatic fingerprint recognition systems, such as AFIS (Automated Fingerprint Identification System). This work proposes a new method for fingerprint image quality estimation using local features (contrast, curvature and ridge flow). A fuzzy inference system is used to combine these features into a single image quality score. Tests are carried out with fingerprint images from the Fingerprint Verification Competition 2006 (FVC 2006) DB2-A database and fingerprint matching software BOZORTH3 from NIST Biometric Image Software (NBIS). After the removal of 5%, 10% and 15% of the poorest quality fingerprints from the DB2-A, we obtained an improvement of 30.6%, 32.6% and 37.9% in EER (Equal Error Rate), respectively.
KEYWORDS
Fingerprint Image Quality, Fuzzy Inference Systems, Digital Image Processing.
