Page 165 - 《应用声学》2025年第1期

第 44 卷 第 1 期 袁博等： 信号分离在深海定位中的应用 161

[3] Hershey J R, Chen Z, Roux J L. Deep clustering: Discrim- [7] Luo Y, Mesgarani N. TasNet: Time-domain audio separa-
inative embeddings for segmentation and separation[C]. tion network for real-time, single-channel speech separa-
2016 IEEE International Conference on Acoustics, Speech tion[C]. 2018 IEEE International Conference on Acoustics,
and Signal Processing (ICASSP), 2016: 31–35. Speech and Signal Processing (ICASSP), 2018: 696–700.
[4] Chen Z, Luo Y, Mesgarani N. Deep attractor network for [8] 成帅, 张海剑, 孙洪. 结合时变滤波和时频掩码的语音增强方
single-microphone speaker separation[C]. 2017 IEEE In- 法 [J]. 信号处理, 2019, 35(4): 601–608.
ternational Conference on Acoustics, Speech and Signal Cheng Shuai, Zhang Haijian, Sun Hong. Joint time-
Processing (ICASSP), 2017: 246–250. varying filtering and masking for speech enhancement[J].
[5] Kolbaek M, Yu D, Tan Z H. Multitalker speech sepa- Journal of Signal Processing, 2019, 35(4): 601–608.
ration with utterance-level permutation invariant train- [9] Pariente M, Cornell S, Deleforge A, et al. Filterbank de-
ing of deep recurrent neural networks[J]. IEEE/ACM sign for end-to-end speech separation[C]. IEEE Interna-
Transactions on Audio, Speech and Language Processing tional Conference on Acoustics, Speech, and Signal Pro-
(TASLP), 2017, 25(10): 1901–1913. cessing, 2020: 6359–6363.
[6] Luo Y, Chen Z, Yoshioka T. Dual-Path RNN: Efficient [10] Luo Y, Mesgarani N. Conv-TasNet: Surpassing ideal
long sequence modeling for time-domain single-channel time–frequency magnitude masking for speech separa-
speech separation[C]. ICASSP 2020- 2020 IEEE Interna- tion[J]. IEEE/ACM Transactions on Audio, Speech, and
tional Conference on Acoustics, Speech and Signal Pro- Language Processing, 2019: 27(8): 1256–1266.
cessing (ICASSP), 2020: 46–50.