Double adversarial networks for monaural speech enhancement

This project implements our paper "Double adversarial networks for monaural speech enhancement". To reproduce the WER results in our paper, please install the newest Kaldi tookits and obtain the CHiME-2 dataset firstly.

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
enhancement_targets		enhancement_targets
loss_functions		loss_functions
models		models
modules		modules
utils		utils
wav_scp		wav_scp
.gitignore		.gitignore
README.md		README.md
chime2_calc_target_sound.py		chime2_calc_target_sound.py
chime2_sph2wav.py		chime2_sph2wav.py
dnn_gan_lm_mask.py		dnn_gan_lm_mask.py
enhance_feat_to_file.py		enhance_feat_to_file.py
enhance_fft_to_file.py		enhance_fft_to_file.py
enhance_ideal_fft_to_file.py		enhance_ideal_fft_to_file.py
feature_extractor.py		feature_extractor.py
kaldi_fbank_dataset.py		kaldi_fbank_dataset.py
kaldi_fbank_extractor.py		kaldi_fbank_extractor.py
kaldi_fft_dataset.py		kaldi_fft_dataset.py
kaldi_helper.py		kaldi_helper.py
segan_dataset.py		segan_dataset.py
speech_utils.py		speech_utils.py
test.py		test.py
test_segan.py		test_segan.py
train.py		train.py
train_adv_fmse.py		train_adv_fmse.py
train_adversarial.py		train_adversarial.py
train_cogan.py		train_cogan.py
train_conditional_adversarial.py		train_conditional_adversarial.py
train_double_adversarial.py		train_double_adversarial.py
train_double_adversarial_dist2adv.py		train_double_adversarial_dist2adv.py
train_fft_domain.py		train_fft_domain.py
train_fmse.py		train_fmse.py
train_gan.py		train_gan.py
train_mel_domain.py		train_mel_domain.py
train_regular.py		train_regular.py
train_regular_dist2adv.py		train_regular_dist2adv.py
train_regular_dyn_alpha.py		train_regular_dyn_alpha.py
train_segan.py		train_segan.py
train_valid_no_aug.py		train_valid_no_aug.py
train_wang2018inves.py		train_wang2018inves.py

ZhihaoDU/du2020dan

Folders and files

Latest commit

History

Repository files navigation

Double adversarial networks for monaural speech enhancement

About

Resources

Stars

Watchers

Forks

Languages