Table 2. Optimization results of seven cases compared to radiologist data by median values.
| case number | REGION GROWING | ADAPTIVE THRESHOLD | ACTIVE CONTOURS | |||
| SC | SG | SC | SG | SC | SG | |
| 20091223 | 95.6% | 79.9% | 92.0% | 92.0% | 94.0% | 82.6% |
| 20100930 | 95.9% | 91.8% | 88.8% | 90.0% | 95.3% | 82.5% |
| 20101221 | 90.3% | 85.6% | 88.6% | 87.8% | 91.0% | 86.9% |
| 20110228 | 88.8% | 84.8% | 83.7% | 83.7% | 88.3% | 86.4% |
| 20110421 | 92.6% | 89.2% | 73.6% | 55.4% | 80.2% | 75.6% |
| 20110624 | 94.4% | 94.4% | 82.7% | 82.7% | 91.2% | 88.9% |
| 20110707 | 93.4% | 93.4% | 87.6% | 87.6% | 91.5% | 91.5% |
| median | 89.2% | 87.6% | 86.4% | |||
Presented are median similarities of seven cases segmented using three segmentation algorithms and compared to models that were generated by radiologist manual segmentation. Every similarity (S) is the median value of similarities from each slice of a case, and was evaluated using individual, case-specific parameters (SC) or using globally-optimized parameters (SG).