Table 4.
Statistical evaluation super-clustering approach for fully automated single particle picking and EMAN2 [29] performances using a beta-galactosidase image [30]. The table reports True Positive (TP) picking results where the correct particles are picked, False Negative (FN) picking results where some good particles are missed, and False Positive (FP) picking results where the incorrect particles (other objects such as background or artificial objects) are picked as particles.
| Evaluation Metric | Our Approach | EMAN2 |
|---|---|---|
| Total Particles Number | 579 | 579 |
| True Positive (TP) | 574 | 101 |
| False Negative (FN) | 5 | 478 |
| False Positive (FP) | 9 | 0 |
| Sensitivity/Recall (%) | 0.99136442 | 0.174439 |
| Precision (%) | 0.98456261 | 1 |
| Accuracy (%) | 0.97619048 | 0.174439 |
| F1 Score (%) | 0.98795181 | 0.297059 |