22 citations · 33 across the 9 of their papers we have counts for
Showing 2023Show all
2 papers · 1 filter
eess.AS2023★ 22 cited
NeXt-TDNN: Modernizing Multi-Scale Temporal Convolution Backbone for Speaker Verification
Hyun-Jun Heo, Ui-Hyeop Shin, Ran Lee +2
In speaker verification, ECAPA-TDNN has shown remarkable improvement by utilizing one-dimensional(1D) Res2Net block and squeeze-and-excitation(SE) module, along with multi-layer fe…
eess.AS2023
Statistical Beamformer Exploiting Non-stationarity and Sparsity with Spatially Constrained ICA for Robust Speech Recognition
Ui-Hyeop Shin, Hyung-Min Park
In this paper, we present a statistical beamforming algorithm as a pre-processing step for robust automatic speech recognition (ASR). By modeling the target speech as a non-station…