Table 2.
Evaluation results of positive unanimous voting (PUV), two-stage Bayesian classifier [3], and surface-based linear discriminant analysis using F-score, true positive rate, and false positive rate.
| F-score | TPR | FPR | ||||
|---|---|---|---|---|---|---|
| Mean | SD | Mean | SD | Mean | SD | |
| TSBC | 0.0708 | 0.1131 | 0.3933 | 0.2754 | 0.0536 | 0.0270 |
| SLDA | 0.2060 | 0.1454 | 0.5374 | 0.2979 | 0.0212 | 0.0241 |
| PUV1 | 0.1251 | 0.1324 | 0.3008 | 0.1657 | 0.0160 | 0.0091 |
| PUV2 | 0.3039 | 0.1865 | 0.2377 | 0.1801 | 0.0211 | 0.0022 |
PUV1 and PUV2 represent the results before and after the connected region size analysis step, respectively.