Table 2 Comparison of different CNN and RNN pair
From: A multi-label classification model for full slice brain computerised tomography image
Model | Parameters | Precision | Recall | F1 |
|---|---|---|---|---|
ResNet50-GRU | 10,106,889 | 63.56% | 47.11% | 0.5411 |
ResNet50-LSTM | 13,253,641 | 57.35% | 49.16% | 0.5259 |
VGG16-GRU | 5,382,153 | 67.57% | 61.04% | 0.6412 |
VGG16-LSTM | 6,956,041 | 58.13% | 57.87% | 0.5794 |
VGG19-GRU | 5,382,153 | 58.61% | 57.49% | 0.5802 |
VGG19-LSTM | 6,956,041 | 46.89% | 65.45% | 0.5462 |
DenseNet121-GRU | 6,957,065 | 60.93% | 45.24% | 0.5168 |
DenseNet121-LSTM | 9,055,241 | 48.59% | 47.62% | 0.4883 |