Fingerprint and On-Line Signature Verification Competitions at ICB 2009

Abstract

Abstract. This paper describes the objectives, the tasks proposed to the participants and the associated protocols in terms of database and assessment tools of two present competitions on fingerprints and on-line signatures, the results of which will be ready for presentation at the next ICB conference. The particularity of the fingerprint competition is to be an on-line competition, for evaluation of fingerprint verification tools such as minutiae extractors and matchers as well as complete systems. The on-line signature competition will test the influence of multi-session, environmental conditions (still and mobility) and signature complexity on the performance of complete systems using two datasets extracted from the BioSecure database.

Impact of time variability in off-line writer identification and verification

“…The results of this paper highlight the importance of this phenomenon and encourage its consideration in future technology benchmarks, e.g. [18,19]. Finally, the results of this paper motivates us to study the individual factors that make some signatures and writers to be more consistent in time than others, in order to develop quality measures that can predict the verification/identification performance [20].…”

Continuous Biometric Verification for Non-Repudiation of Remote Services

“… Sys 1 with Google FaceNet [32], for face recognition, which showed at FAR 10 -6 a True Acceptance Rate (TAR) of 86.473% in Megaface challenge [33], and the impressive accuracy result of 99.63%±0.09 in LFW benchmark [34];  Sys 2 with IRITECH algorithm for iris recognition, which showed at FAR 10 -6 a FRR of 0.002 in the context of NIST Iris Exchange IREX I [35];  Sys 3 with Neurotechnology algorithm for fingerprint recognition, which had at FAR≤ 10 −2 a FRR of 0.083 in the FVC-onGoing online evaluation [36].…”

Introducing an Experimental Framework in C# for Fingerprint Recognition

“…d) The users do not have control over the fingerprint databases and they cannot create an experiment with a custom protocol for performance evaluation. For instance, latent fingerprint identification is a very active research topic at present but users cannot use their own databases with the "FVC-onGoing web-based automated evaluation system" [8] because the system does not allow the users to upload custom databases. e) The tools do not contain any protocol for performance evaluation.…”

“…For example, to test "SourceAFIS SDK" [31] in FVC2004 [17] databases, the users have to implement the performance evaluation protocol. [29] X X FVC-onGoing web-based automated evaluation system [8] X X X SourceAFIS SDK [31] X X MCC SDK [4,5,10] X X X VeriFinger SDK [23] X X X X Fingerprint SDK [12] X X X X BiometricSDK [27] X X IDKit PC SDK [14] X X X X To deal with all these limitations, this paper introduces a fingerprint recognition framework containing matching algorithms, feature extraction algorithms, and experimental protocols. This framework is implemented in .Net Framework using the C# language for the following reasons:…”

“…This framework was used to create a new version of M3gl [20] which is among the top ten algorithms according to the accuracy indicators in database FMISO-HARD of FVC-onGoing [8].…”