A speaker verification backend with robust performance across conditions

L Ferrer, M McLaren, N Brümmer - Computer Speech & Language, 2022 - Elsevier
In this paper, we address the problem of speaker verification in conditions unseen or
unknown during development. A standard method for speaker verification consists of …

How to make embeddings suitable for PLDA

Z Li, R Xiao, H Chen, Z Zhao, W Wang… - Computer Speech & …, 2023 - Elsevier
Probabilistic linear discriminant analysis (PLDA) is widely implemented in speaker
verification tasks. However, PLDA has limitations owing to its assumptions. In this study, we …

PLDA inspired Siamese networks for speaker verification

S Ramoji, P Krishnan, S Ganapathy - Computer Speech & Language, 2022 - Elsevier
The deep learning methodologies in state-of-the-art speaker recognition systems are
predominantly limited to the extraction of recording level embeddings. This is usually …

Fusing linguistic and acoustic information for automated forensic speaker comparison

EK Sergidou, R Ypma, J Rohdin, M Worring, Z Geradts… - Science & Justice, 2024 - Elsevier
Verifying the speaker of a speech fragment can be crucial in attributing a crime to a suspect.
The question can be addressed given disputed and reference speech material, adopting the …

[HTML][HTML] Joint speaker encoder and neural back-end model for fully end-to-end automatic speaker verification with multiple enrollment utterances

C Zeng, X Miao, X Wang, E Cooper… - Computer Speech & …, 2024 - Elsevier
Conventional automatic speaker verification systems can usually be decomposed into a
front-end model such as time delay neural network (TDNN) for extracting speaker …

Graph Neural Network Backend for Speaker Recognition

L He, R Li, M Niu - arXiv preprint arXiv:2308.08767, 2023 - arxiv.org
Currently, most speaker recognition backends, such as cosine, linear discriminant analysis
(LDA), or probabilistic linear discriminant analysis (PLDA), make decisions by calculating …

[PDF][PDF] Supervised Learning Approaches for Language and Speaker Recognition

S Ramoji - 2023 - leap.ee.iisc.ac.in
In the age of artificial intelligence, it is important for machines to figure out who is speaking
automatically and in what language-a task humans are naturally capable of. Developing …

Coupling a generative model with a discriminative learning framework for speaker verification

X Lu, P Shen, Y Tsao, H Kawai - IEEE/ACM Transactions on …, 2021 - ieeexplore.ieee.org
The task of speaker verification (SV) is to decide whether an utterance is spoken by a target
or an imposter speaker. In most studies of SV, a log-likelihood ratio (LLR) score is estimated …

話者認識システムとなりすまし対策――競争型評価ワークショップからみる分野の潮流と最新手法――

俵直弘 - 日本音響学会誌, 2022 - jstage.jst.go.jp
分解することでパラメタ数を減らした factorized TDNN [6] などが提案されている. また, TDNN
の各層を階層的な残差 (residual-like connection in residual; Res2) ブロックと SE (squeeze …

[PDF][PDF] Spoofing-aware Speaker Verification System Robust Against Domain and Channel Mismatches (ドメインとチャンネルのミスマッチに頑健ななりすまし対応話者照合 …

ソウチョウ - ir.soken.ac.jp
Automatic speaker verification (ASV) has shown immense potential across domains like
security, forensic analysis, and human-computer interaction. However, real-world …