Research Article
Assisted Diagnosis of Alzheimer’s Disease Based on Deep Learning and Multimodal Feature Fusion
Table 2
Experimental results of sMRI data (%).
| | AD versus NC | Acc | Sen | Spec | Precision | Recall | f1-score | AUC |
| | ResNet_10 | 80.4 | 77.0 | 82.7 | 74.9 | 77.0 | 75.8 | 80.9 | | DenseNet_121 | 74.0 | 53.0 | 88.0 | 77.1 | 53.0 | 61.0 | 79.5 | | 3DShuffleNet | 85.2 | 69.0 | 96.0 | 93.3 | 69.0 | 79.0 | 86.9 |
| | AD versus MCI | Acc | Sen | Spec | Precision | Recall | f1-score | AUC | | ResNet_10 | 77.5 | 77.0 | 78.0 | 78.0 | 77.0 | 77.4 | 86.2 | | DenseNet_121 | 62.0 | 61.0 | 63.0 | 64.2 | 61.0 | 61.1 | 68.3 | | 3DShuffleNet | 84.0 | 84.0 | 84.0 | 84.9 | 84.0 | 84.1 | 91.9 |
| | MCI versus NC | Acc | Sen | Spec | Precision | Recall | f1-score | AUC | | ResNet_10 | 64.8 | 38.0 | 82.7 | 61.6 | 38.0 | 45.8 | 65.5 | | DenseNet_121 | 53.6 | 52.0 | 54.6 | 47.5 | 52.0 | 47.1 | 55.5 | | 3DShuffleNet | 64.8 | 43.0 | 79.3 | 60.2 | 43.0 | 48.0 | 67.3 |
| | EMCI versus LMCI | Acc | Sen | Spec | Precision | Recall | f1-score | AUC | | ResNet_10 | 54.0 | 56.0 | 52.0 | 53.7 | 56.0 | 54.1 | 65.2 | | DenseNet_121 | 54.0 | 28.0 | 80.0 | 62.5 | 28.0 | 36.7 | 48.2 | | 3DShuffleNet | 53.0 | 40.0 | 66.0 | 56.7 | 40.0 | 45.6 | 58.0 |
|
|
