Table 2.
Results of round-robin analyses for all classification and view selection protocols given as the area under the ROC curve.
Classification Protocol
|
|||
---|---|---|---|
View Protocol | A | B | C |
All images | 0.86
[0.813; 0.895] |
0.87
[0.827; 0.905] |
0.83
[0.758; 0.861] |
| |||
Random images | 0.85
[0.806; 0.889] |
0.85
[0.810; 0.892] |
0.83
[0.745; 0.870] |
| |||
First image only | 0.86
[0.811; 0.893] |
0.86
[0.816; 0.898] |
0.86
[0.816; 0.898] |
Note – Data are indicated as AUC [two-sided 95% confidence interval].
Protocol A: Feature averaging
Protocol B: Classifier output averaging
Protocol C: No averaging