A Hybrid Continuity Loss to Reduce Over-Suppression for Time-domain Target Speaker Extraction.

Zexu Pan Meng Ge Haizhou Li

Published in: CoRR (2022)

Keyphrases

frequency domain
information extraction
hybrid approaches
significantly reduced
computer vision
neural network
artificial intelligence
artificial neural networks
multiresolution
databases
information retrieval
infrared
speech recognition
machine learning
hybrid learning
target recognition
speaker recognition