Research Article
Automated Detection Model Based on Deep Learning for Knee Joint Motion Injury due to Martial Arts
Table 2
Identification accuracy of different models.
| Injured part | Sensitivity | Specificity | F1 value | Mean accuracy |
| Meniscus | 95.5/93.7 | 61.5/60.1 | 74.9/73.5 | 83.2/80.3 | Tendon | 91.3/91.3 | 87.3/87.2 | 89.2/89.1 | 89.0/89.0 | Ligamentous | 91.1/89.8 | 83.5/82.5 | 87.0/85.6 | 88.0/85.7 | Bone and apical cartilage | 88.4/87.5 | 87.6/86.7 | 88.0/87.1 | 85.9/85.0 | Synovial capsule | 91.2/89.4 | 74.7/73.5 | 82.3/80.7 | 85.6/82.5 | Soft tissue | 86.1/85.2 | 85.8/84.8 | 85.9/84.9 | 83.5/82.3 |
|
|