Research Article
MAM: Multiple Attention Mechanism Neural Networks for Cross-Age Face Recognition
Table 3
The ACC of the improved model with residual-attention mechanism in different deployment modes on the CACD-VS and MORPH datasets.
| | Model | ACC on CACD-VS | ACC on MORPH |
| | ResNet50(fine-tuned by MORPH) | 95.31% | 95.83% | | Residual attention ResNet50 (stage 2) | 98.01% | 98.10% | | Residual attention ResNet50 (stage 3) | 97.82% | 97.97% | | Residual attention ResNet50 (stage 4) | 97.21% | 97.41% | | Residual attention ResNet50 (stage 5) | 96.58% | 96.69% | | Residual attention ResNet50 (stage 2, 3) | 98.06% | 98.14% | | Residual attention ResNet50 (stage 2, 4) | 97.99% | 98.17% | | Residual attention ResNet50 (stage 2-4) | 98.12% | 98.23% | | Residual attention ResNet50 (stage 2, 5) | 97.87% | 98.15% | | Residual attention ResNet50 (stage 2-5) | 98.12% | 98.22% | | Residual attention ResNet50 (stage 3, 4) | 97.91% | 97.99% | | Residual attention ResNet50 (stage 3, 5) | 97.91% | 98.00% | | Residual attention ResNet50 (stage 3-5) | 97.93% | 98.02% | | Residual attention ResNet50 (stage 4, 5) | 97.28% | 97.46% |
|
|
