Implementation of low-latency electrolaryngeal speech enhancement based on multi-task CLDNN.

Kazuhiro Kobayashi Tomoki Toda

Published in: EUSIPCO (2020)

Keyphrases

multi task
low latency
multi task learning
learning tasks
speech enhancement
high throughput
high speed
noisy environments
multi class
feature selection
signal to noise ratio
machine learning
real time
highly efficient
data streams
learning algorithm
speech signal
data sets
single channel