Print Email Facebook Twitter Hardware and algorithm study for a fingerprint verification system Title Hardware and algorithm study for a fingerprint verification system Author Mandado, Y. Contributor Van Leuken, T.G.R. (mentor) Faculty Electrical Engineering, Mathematics and Computer Science Department CAS Programme Computer Engineering Date 2012-03-15 Abstract Biometrics is the area of knowledge comprising the methods for unique human recognition, based upon one or more intrinsic physical or behavioural traits. The term has its origin in the ancient Greek language (bios ="life" plus metron ="measure"). Among all the biometric techniques[1], some of them of recent development, fingerprinting has been the most widely used during the 20th Century. The maturity of this biometric technique and, mainly, the dramatic improvement of the capture devices have lead to the proposal of fingerprinting in multiple applications. In this thesis a hardware architecture for a Fingerprint Verification System (FVS) in a mobile embedded device, one of the most challenging problems in this field, is developed. The objective of the system is to decide whether an input fingerprint belongs to the same individual than a previously stored one. This procedure is called verification, in contrast with identification, that tries to decide which fingerprint in a database shows higher resemblances with an input one. In this thesis, new algorithms are proposed for fingerprint matching, taking as a departure point literature and research publications from last decades. The main development is the introduction of Delta and Core detection algorithms. With this points a classification process of the fingerprints is implemented. To reference this document use: http://resolver.tudelft.nl/uuid:5b3a8ae9-a605-4933-adcb-b8f736425423 Embargo date 2012-02-24 Part of collection Student theses Document type master thesis Rights (c) 2012 Mandado, Y. Files PDF Thesis.pdf 1.14 MB Close viewer /islandora/object/uuid:5b3a8ae9-a605-4933-adcb-b8f736425423/datastream/OBJ/view