A Siamese MLP that fuses speaker embeddings from noisy and DeepFilterNet-enhanced speech cuts speaker verification error at SNR -10 dB and below, while degrading performance near 0 dB.
The V oxCeleb1 training set comprises 148,642 utterances from 1,211 speakers, while the test set contains 4,874 utterances from 40 speakers
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
A Framework for Robust Speaker Verification in Highly Noisy Environments Leveraging Both Noisy and Enhanced Audio
A Siamese MLP that fuses speaker embeddings from noisy and DeepFilterNet-enhanced speech cuts speaker verification error at SNR -10 dB and below, while degrading performance near 0 dB.