A
Albert Swart
Researcher at Stellenbosch University
Publications - 13
Citations - 260
Albert Swart is an academic researcher from Stellenbosch University. The author has contributed to research in topics: Speaker recognition & Calibration (statistics). The author has an hindex of 6, co-authored 11 publications receiving 242 citations.
Papers
More filters
Proceedings ArticleDOI
The RedDots Data Collection for Speaker Recognition
Kong Aik Lee,Anthony Larcher,Guangsen Wang,Patrick Kenny,Niko Brümmer,David A. van Leeuwen,Hagai Aronowitz,Marcel Kockmann,Carlos Vaquero,Bin Ma,Haizhou Li,Themos Stafylakis,Md. Jahangir Alam,Albert Swart,Javier Pérez +14 more
TL;DR: This paper describes data collection efforts conducted as part of the RedDots project which is dedicated to the study of speaker recognition under conditions where test utterances are of short duration and of variable phonetic content.
A comparison of linear and non-linear calibrations for speaker recognition
TL;DR: The authors generalize the linear recipes to non-linear ones and experiment with a nonlinear, non-parametric, discriminative PAV solution, as well as parametric, generative, maximum-likelihood solutions that use Gaussian, Student's T and normal-inverse-Gaussian score distributions.
Proceedings ArticleDOI
Bayesian calibration for forensic evidence reporting
Niko Brümmer,Albert Swart +1 more
TL;DR: In this paper, a Bayesian likelihood-ratio distribution is used to report the weight of evidence for forensic speaker recognition, where there may be very little background material for estimating score calibration parameters.
Posted Content
Bayesian calibration for forensic evidence reporting
Niko Brümmer,Albert Swart +1 more
TL;DR: A Bayesian solution for the problem in forensic speaker recognition, where there may be very little background material for estimating score calibration parameters, is introduced and developed, which results in a Bayesian likelihood-ratio as the vehicle for reporting weight of evidence.
Proceedings ArticleDOI
Analysis and Description of ABC Submission to NIST SRE 2016.
Oldřich Plchot,Pavel Matějka,Anna Silnova,Ondřej Novotný,Mireia Diez Sánchez,Johan Rohdin,Ondřej Glembek,Niko Brümmer,Albert Swart,Jesús Jorrín-Prieto,Paola García,Luis Buera,Patrick Kenny,Jahangir Alam,Gautam Bhattacharya +14 more
TL;DR: It is shown that testing on mismatched, non-English and short duration data introduced in NIST SRE 2016 is a difficult problem for current state-of-theart systems.