Research Article
[Retracted] Research on Chest Disease Recognition Based on Deep Hierarchical Learning Algorithm
Table 5
The trainable parameters and FLOPs of ResNet-101-GCN, SGGCN-101, and MobileNetV2-GCN.
| Structure | Trainable parameters | FLOPs |
| ResNet-101-GCN | 47,308,864 | 16,017,450,516 | SGGCN-101 | 12,427,684 (−73.73%) | 3,072,345,732 (−80.82%) | MobileNetV2-GCN | 5,459,712 (−88.46%) | 661,395,120 (−95.88%) |
|
|