Table 4 Kappa value between proposed AI model and expert prediction with fetal’s heart and children’s heart US images
From: Automatic echocardiographic anomalies interpretation using a stacked residual-dense network model
| Â | Â | Â | Actual condition | Â | Kappa value |
|---|---|---|---|---|---|
| Â | Â | Â | Cardiac septal defect | Normal | Â |
Fetomaternal cardiologist prediction | Expert 1 | Cardiac septal defect | 765 (99.73%) | 2 (0.27%) | 0.912 |
| Â | Â | Normal | 69 (8.26%) | 766 (91.74%) | Â |
| Â | Expert 2 | Cardiac septal defect | 709 (69.10%) | 317 (30.10%) | 0.540 |
| Â | Â | Normal | 59 (10.12%) | 524 (89.88%) | Â |
Pediatric cardiologist prediction | Expert 1 | Cardiac septal defect | 171 (83.01%) | 35 (16.99%) | 0.605 |
| Â | Â | Normal | 50 (22.32%) | 174 (77.68%) | Â |
| Â | Expert 2 | Cardiac septal defect | 101 (74.81%) | 34 (25.19%) | 0.291 |
| Â | Â | Normal | 120 (40.68%) | 175 (59.32%) | Â |
| Â | Expert 3 | Cardiac septal defect | 167 (71.06%) | 68 (28.94%) | 0.431 |
| Â | Â | Normal | 54 (27.69%) | 141 (72.31%) | Â |