Table 5 Accuracy for interchanging training and testing dataset
From: Comparative study of singing voice detection based on deep neural networks and ensemble learning
| Â | Training: Jamndo Testing: FMA (%) | Training: FMA Testing: Jamendo (%) |
|---|---|---|
MCNN | 76.48 | 85.72 |
SCNN | 81.69 | 89.62 |
Capsule net | 70.37 | 73.36 |