Making Likelihood Ratios Digestible for Cross-Application Performance Assessment

Abstract

Performance estimation is crucial to the assessment of novel algorithms and systems. In detection error tradeoff (DET) diagrams, discrimination performance is solely assessed targeting one application, where cross-application performance considers risks resulting from decisions, depending on application constraints. For the purpose of interchangeability of research results across different application constraints, we propose to augment DET curves by depicting systems regarding their support of security and convenience levels. Therefore, application policies are aggregated into levels based on verbal likelihood ratio scales, providing an easy to use concept for business-to-business communication to denote operative thresholds. We supply a reference implementation in Python, an exemplary performance assessment on synthetic score distributions, and a fine-tuning scheme for Bayes decision thresholds, when decision policies are bounded rather than fix. Index Terms-Bayes decision framework, biometric verification, binary decisions, detection error tradeoff (DET), verbal scales.

Dynamic signatures: A review of dynamic feature variation and forensic methodology

“…Additionally, forensic science has gained insight from the findings in biometrics, with several authors developing movement-modeling techniques [60][61][62][63][64] to represent handwriting movement. There has been a research effort towards descriptor development and feature selection in dynamic signatures [9,10,[147][148][149][150] or adapting automated comparison systems to forensic purposes [151][152][153][154]. Furthermore, the criticism has led forensic scientists to provide evidence for expert opinion reliability [155][156][157][158][159][160], to reconsider the identification and unicity paradigms [161,162] and to publish their methods [76] and best practices [163,164].…”

Bayesian evaluation of dynamic signatures in operational conditions

“…There have also been many advances in methods for validation criteria [45][46][47]. Further progress in feature characterization could be achieved through computational techniques [18,19,[48][49][50][51][52][53][54]. Forensic handwriting examination has been able to take advantage of this research [31,[55][56][57], but no major methodological breakthroughs related to dynamic signatures have been achieved in recent years.…”

Privacy-Preserving Indexing of Iris-Codes with Cancelable Bloom Filter-based Search Structures

“…the size of the key space). Figure 4 shows DET curves (with axes using a standard deviate scale [30]) for some of the best performing system configurations. Plots for each IC parts ∈ {2, 4, 8} show the three best configurations in terms of biometric performance and three best configurations in terms of workload reduction.…”