nn-singal-processing-papers

List of NN based singal processing papers

Adaptive Noise Suppression (Speech Enhancement)

Time-Frequency Domain

• DCUnet: Phase-aware speech enhancement with Deep Complex U-Net (SNU, ICLR, 2019)

• DCCRN

  • DCCRN: DCCRN: Deep Complex Convolution Recurrent Network for Phase-Aware Speech Enhancement (NWPU, 2020)
  • DCCRN+: DCCRN+: Channel-wise Subband DCCRN with SNR Estimation for Speech Enhancement (NWPU, 2021)
  • S-DCCRN: S-DCCRN: Super Wide Band DCCRN with learnable complex feature for speech enhancement (NWPU, 2022)
  • Spatial-DCCRN: SPATIAL-DCCRN: DCCRN EQUIPPED WITH FRAME-LEVEL ANGLE FEATURE AND HYBRID FILTERING FOR MULTI-CHANNEL SPEECH ENHANCEMENT (NWPU, 2022)

• DesNet: DESNET: A MULTI-CHANNEL NETWORK FOR SIMULTANEOUS SPEECH DEREVERBERATION, ENHANCEMENT AND SEPARATION (NWPU, 2020)

• PHASEN: PHASEN: A Phase-and-Harmonics-Aware Speech Enhancement Network (USTC, AAAI, 2020)

• DPCRN: DPCRN: Dual-Path Convolution Recurrent Network for Single Channel Speech Enhancement (NJU, Interspeech, 2021)

• BSRNN: High Fidelity Speech Enhancement with Band-split RNN(Tencent, 2022)

• UFormer: Uformer: A Unet based dilated complex & real dual-path conformer network for simultaneous speech enhancement and dereverberation (NWPU, 2022)

Time Domain

• WaveUnet: Wave-U-Net: A Multi-Scale Neural Network for End-to-End Audio Source Separation (QMUL,ISMIR, 2018)
• TCNN: TCNN: TEMPORAL CONVOLUTIONAL NEURAL NETWORK FOR REAL-TIME SPEECH ENHANCEMENT IN THE TIME DOMAIN (Ohio, ICASSP, 2019)
• DP-SARNN: Dual-path Self-Attention RNN for Real-Time Speech Enhancement (Ohio, 2021)

Acoustic Echo Cancellation

• wRLS-DFSMN: Weighted Recursive Least Square Filter and Neural Network based Residual Echo Suppression for the AEC-Challenge (Alibaba, ICASSP, 2021)
• GCCRN Acoustic Echo Cancellation using Deep Complex Neural Network with Nonlinear Magnitude Compression and Phase Information (IACAS, Interspeech, 2021)

Automatic Gain Control

Speech Seperation

TODO: add important models from ESPnet and asteroid.

Single Channel

Multiple Channel

Joint Optimization

• Deep Learning for Joint Acoustic Echo and Noise Cancellation with Nonlinear Distortions (Ohio, Interspeech, 2019)
• Low-Complexity, Real-Time Joint Neural Echo Control and Speech Enhancement Based On PercepNet (Amazon, ICASSP, 2021)
• NN3A: NN3A: Neural Network supported Acoustic Echo Cancellation, Noise Suppression and Automatic Gain Control for Real-Time Communications (Alibaba, ICASSP, 2022)

Masking

• IBM: On Ideal Binary Mask As the Computational Goal of Auditory Scene Analysis (Ohio, 2005)
• IRM: Ideal ratio mask estimation using deep neural networks for robust speech recognition (Ohio, 2013)
• PSM: Phase-Sensitive and Recognition-Boosted Speech Separation using Deep Recurrent Neural Networks (Microsoft, 2015)
• CRM: Complex Ratio Masking for Monaural Speech Separation (Ohio, 2015)