Title | ||
---|---|---|
Noisy-target Training: A Training Strategy for DNN-based Speech Enhancement without Clean Speech |
Abstract | ||
---|---|---|
Deep neural network (DNN)-based speech enhancement ordinarily requires clean speech signals as the training target. However, collecting clean signals is very costly because they must be recorded in a studio. This requirement currently restricts the amount of training data for speech enhancement to less than 1/1000 of that of speech recognition which does not need clean signals. Increasing the amou... |
Year | DOI | Venue |
---|---|---|
2021 | 10.23919/EUSIPCO54536.2021.9616166 | 2021 29th European Signal Processing Conference (EUSIPCO) |
Keywords | DocType | ISSN |
Single-channel speech enhancement,deep neural network (DNN),training target,Noise2Noise | Conference | 2076-1465 |
ISBN | Citations | PageRank |
978-9-0827-9706-0 | 3 | 0.39 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Takuya Fujimura | 1 | 3 | 0.39 |
Koizumi Yuma | 2 | 41 | 11.75 |
Kohei Yatabe | 3 | 16 | 10.36 |
Ryoichi Miyazaki | 4 | 3 | 1.06 |