Fingerprint Quality Assessment: Matching Performance and Image Quality

Zhigang Yao (1) , Jean-Marie Le Bars (2) , Christophe Charrier (2) , Christophe Rosenberger (2)

1 Capital Normal University [Beijing]
2 Equipe SAFE - Laboratoire GREYC - UMR6072

This article chiefly focuses on Fingerprint Quality Assessment (FQA) applied to the Automatic Fingerprint Identification System (AFIS). In our research work, different FQA solutions proposed so far are compared by using several quality metrics selected from the existing studies. The relationship between the quality metric and the matching performance is specifically analyzed via such a comparison and an extra discussion based on the sample utility. This study is achieved via an in-teroperability analysis by using two different matching tools. Experimental results represented by the global Equal Error Rate (EER) further reveal the constraint of the existing quality assessment solutions in predicting the matching performance. Experiments are performed via several public-known fingerprint databases.

Domaines: Cryptographie et sécurité [cs.CR]

Zhigang Yao, Jean-Marie Le Bars, Christophe Charrier, Christophe Rosenberger. Fingerprint Quality Assessment: Matching Performance and Image Quality. Jiang, R., Al-Madeed, S., Bouridane, A., Crookes, D., Beghdadi, A. Biometric Security and Privacy: Opportunities & Challenges in The Big Data Era, Springer, 2016, 978-3-319-47301-7. ⟨hal-01406677⟩