Table 2.
Semantic segmentation performances of L-K detection and ring detection (IOU: intersection over union, BF1: boundary F-1).
| Performance Deleted Extra Space | |||||
|---|---|---|---|---|---|
| Area | Mean Accuracy | Mean IOU | BF1 Score | Dataset | |
| L-K Detection |
Kidney | 0.9682 | 0.8088 | 0.4650 | Training: 1590 Validation: 530 Test: 530 N:Mi:Mo:S = 30:11:10:2 |
| Liver | 0.9487 | 0.7856 | 0.5228 | ||
| Null | 0.9415 | 0.9341 | 0.8002 | ||
| Ring Detection |
Kidney | 0.8642 | 0.6665 | 0.5510 | |
| Liver | 0.8318 | 0.6576 | 0.5785 | ||
| Null | 0.8318 | 0.8307 | 0.8663 | ||
| Training optimizer: Adam, Minibatch size: 8, Max epoch: 10, Learning rate of 0.001 with decay factor of 0.9, Termination condition: validation accuracy < 0.98 | |||||