Table 1. Cluster accuracies for microsphere phantoms for the different metrics combined with a K-Means algorithm and a Self Organizing Map for two clusters.
|
|
Phantom 1
|
Phantom 2
|
Phantom 3
|
Phantom 4
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| K-Means + | Accuracy | Sensitivity | Specificity | Accuracy | Sensitivity | Specificity | Accuracy | Sensitivity | Specificity | Accuracy | Sensitivity | Specificity | |
| ACF | 62.4 | 98.9 | 41.9 | 76.7 | 85.7 | 71.3 | 69.3 | 88.0 | 62.9 | 87.4 | 99.4 | 80.1 | |
| PHA | 81.4 | 99.8 | 59.6 | 89.5 | 92.1 | 87.3 | 95.6 | 99.9 | 92.1 | 94.0 | 96.8 | 91.6 | |
| SUB | 88.7 | 99.8 | 70.9 | 92.3 | 92.5 | 92.1 | 86.2 | 96.1 | 79.8 | 93.4 | 96.2 | 90.9 | |
| COM | 90.7 | 99.5 | 75.1 | 92.4 | 93.0 | 91.8 | 91.0 | 93.7 | 88.6 | 95.1 | 96.4 | 93.9 | |
| PCA | 87.4 | 99.7 | 68.6 | 93.4 | 94.2 | 92.6 | 95.4 | 99.9 | 91.7 | 98.6 | 99.9 | 97.4 | |
| PCA + SOM | 79.4 ± 1.6 | 100 ± 0.0 | 57.0 ± 1.7 | 90.1 ± 0.7 | 92.4 ± 0.5 | 88.0 ± 1.5 | 91.6 ± 1.5 | 98.4 ± 0.8 | 86.5 ± 2.4 | 98.6 ± 0.5 | 99.4 ± 0.2 | 97.9 ± 1.1 |
|