Forensic speaker recognition: A new method based on extracting accent and language information from short utterances

S Saleem, F Subhan, N Naseer, A Bais… - Forensic Science …, 2020 - Elsevier
… as follows:(1) ∧ c ( Y ) = l o g p ( Y | λ c ) − l o g p ( Y | λ u b m ) (2) c ˆ = arg max c = 1 N ∧
c where Y is a sequence of MFCC features of a test speech sample. Accent/language/…

A comprehensive review on speaker recognition

B Saritha, MA Laskar, RH Laskar - … in Speech and Music Technology …, 2022 - Springer
speaker identification and speaker verification. The process of identifying an unknown speaker
from a set of known speakers is speaker … The acoustic feature sequence is obtained from …

A survey of speaker recognition: Fundamental theories, recognition methods and opportunities

MM Kabir, MF Mridha, J Shin, I Jahan, AQ Ohi - IEEE Access, 2021 - ieeexplore.ieee.org
… The study showed that the speaker trait is primarily a deterministic short-time feature rather
… that aggregated the variable-length input sequence into an utterance level representation. …

Recognition of heavily accented and emotional speech of English and Czech Holocaust survivors using various DNN architectures

JV Psutka, A Pražák, J Vaněk - International Conference on Speech and …, 2021 - Springer
sequence as the objective function to train it. This training procedure (LF-MMI) uses a
sequence … it is possible to use automatic speech recognition technology to search for relevant …

[HTML][HTML] Using a small amount of text-independent speech data for a BiLSTM large-scale speaker identification approach

MK Nammous, K Saeed, P Kobojek - Journal of King Saud University …, 2022 - Elsevier
… of the sequence. For any given timestep t c in a … Speaker Identification task. The dataset
was manually prepared in three languages: Arabic, English, and Polish for 24 speakers with a bit …

[PDF][PDF] Spine2Net: SpineNet with Res2Net and Time-Squeeze-and-Excitation Blocks for Speaker Recognition.

M Rybicka, J Villalba, P Zelasko, N Dehak… - Interspeech, 2021 - isca-archive.org
… In both models, the sequence of the main building blocks is similar. The input feature map
is … The residual part of the ResNet-34 exhibits the same block sequence, with the difference …

Automatic speech recognition system for tonal languages: State-of-the-art survey

J Kaur, A Singh, V Kadyan - Archives of Computational Methods in …, 2021 - Springer
… In this paper a systematic survey on Automatic Speech Recognition (ASR) for tonal …
signal and classify it as the last learned pattern in a sequence. The performance as well as …

Online Speaker Diarization Using Optimized SE-ResNet Architecture

F Kynych, J Zdansky, P Cerva, L Mateju - … Conference on Text, Speech …, 2023 - Springer
… and produces a sequence of speaker embeddings on its … a built-in voice activity detection
module based on a single-… compare the results on the speaker verification task of the original …

[HTML][HTML] Ksponspeech: Korean spontaneous speech corpus for automatic speech recognition

JU Bang, S Yun, SH Kim, MY Choi, MK Lee, YJ Kim… - Applied Sciences, 2020 - mdpi.com
… -to-end speech recognition model trained with KsponSpeech. … be considered in spontaneous
speech recognition in Korean. … -art speech recognition with sequence-to-sequence models. …

Discovering phonetic inventories with crosslingual automatic speech recognition

P Żelasko, S Feng, LM Velazquez, A Abavisani… - Computer Speech & …, 2022 - Elsevier
The high cost of data acquisition makes Automatic Speech Recognition (ASR) model training
problematic for most existing languages, including languages that do not even have a …