Table 5 Performance comparison on LITIS ROUEN dataset
From: Discriminative frequency filter banks learning with neural networks
Method | Accuracy | F-measure | MCC | Error |
|---|---|---|---|---|
TriFB | 96.24 | 96.19 | 96.01 | 3.76 |
GaussFB | 96.33 | 96.44 | 96.11 | 3.67 |
TriFB-DN | 96.61 | 96.50 | 96.39 | 3.39 |
GaussFB-DN | 96.83 | 96.71 | 96.63 | 3.17 |
BandPosFB-DN | 96.84 | 96.71 | 96.64 | 3.16 |
PosFB-DN | 96.15 | 96.04 | 95.91 | 3.85 |
CNN-Gam [9] | 95.8 | 95.8 | – | 4.2 |
CNN-MFCC [9] | 94.0 | 93.7 | – | 6.0 |
CNN-Log [9] | 95.1 | 95.0 | – | 4.9 |
RNN-Gam [8] | 96.4 | 96.6 | – | 3.6 |
RNN-MFCC [8] | 95.4 | 95.8 | – | 4.6 |
RNN-Log [8] | 95.9 | 96.2 | – | 4.1 |