Utilization of exponentiated amplitude spectrum for speech enhancement in highly noisy environments.

Emiko Abe Tetsuya Shimamura

Published in: MWSCAS (2015)

Keyphrases

edge detection
speech enhancement
noise reduction
noisy environments
signal to noise ratio
spectral subtraction
edge preserving
multiscale
single channel
background noise
neural network
noisy speech
speech recognition
speaker identification
multi channel
vocal tract
multiresolution