IIFC-Net: A Monaural Speech Enhancement Network with High-Order Information Interaction and Feature Calibration
Demo: Visit our [demo website] (https://wenbingwei.github.io) for audio samples.
Denoising performance comparison with other systems on the VoiceBank+DEMAND dataset
| Model | Par .(M) | MACs (G/s) | SISNR | PESQ | STOI (%) | CSIG | CBAK | COVL |
|---|---|---|---|---|---|---|---|---|
| Noisy | —— | —— | 8.45 | 1.97 | 92 | 3.34 | 2.45 | 2.63 |
| PHASEN | 8.76 | —— | —— | 2.99 | —— | 4.21 | 3.55 | 3.62 |
| DCCRN | 3.67 | 14.13 | 19.13 | 2.57 | 94 | 3.93 | 2.9 | 3.21 |
| TSTNN | 0.92 | —— | 18.82 | 2.96 | 95 | 4.33 | 3.53 | 3.67 |
| SADNUNet | 2.63 | —— | —— | 2.82 | —— | 4.18 | 3.47 | 3.51 |
| FAF-Net | 6.9 | 108.96 | —— | 3.19 | 95 | 4.13 | 3.38 | 3.66 |
| PFRNet | 4.61 | —— | —— | 3.24 | 95 | 4.48 | 3.7 | 3.9 |
| D2Net | 1.13 | 36.21 | 19.78 | 3.27 | 96 | 4.63 | 3.18 | 3.92 |
| IIFC-Net | 0.586 | 14.91 | 19.82 | 3.28 | 96 | 4.52 | 3.72 | 3.92 |
[1]Wei W, Hu Y, Huang H, He L. IIFC-Net: A Monaural Speech Enhancement Network with High-Order Information Interaction and Feature Calibration. IEEE Signal Processing Letters.SPL, Vol. 31, 2024.