Table 11 State-of-the-art comparison with spectrofusionnet method.
From: SpectroFusionNet a CNN approach utilizing spectrogram fusion for electric guitar play recognition
State-of-the art | Â | Accuracy (%) | F1-score (%) |
|---|---|---|---|
Guitar aataset | |||
 Deep Salience multiple-f0 estimation algorithm | 46 | 90 | |
 SpectroFusionNet-Early Fusion |  | 85.09 | 91.32 |
 SpectroFusionNet-Late Fusion-simple concat |  | 89.47 | 92.71 |
 SpectroFusionNet-Late Fusion-max voting |  | 90.35 | 96.1 |
 SpectroFusionNet-Late Fusion-weighted average |  | 88.60 | 92.9 |
Guitar Style Dataset | |||
 SVM | 84.20 | 81.9 | |
 CNN | 81.10 | 83.1 | |
 SpectroFusionNet-Early Fusion |  | 97.14 | 97.27 |
 SpectroFusionNet-Late Fusion-simple concat |  | 99.12 | 99 |
 SpectroFusionNet-Late Fusion-max voting |  | 100 | 100 |
 SpectroFusionNet-Late Fusion-weighted average |  | 100 | 99.11 |
