Table 2 Table showing comparisons among several feature pre-selection scenarios in terms of run time and efficiency in detection of glial cells for two different validation datasets (VD 1 & 2)
Dataset | Pre-selection | Pre-processing | Data generation | Training | Detected/Actual cells |
|---|---|---|---|---|---|
VD1 | All features | None | – | 11 min 56s | 0/5 |
VD2 | All features | None | – | 2 min 19s | 0/3 |
VD1 | All features | sc-SynO | 4 min 3s | 28 min 24s | 2423/5 |
VD2 | All features | sc-SynO | 4 min 19s | 5 min 8s | 299/3 |
VD1 | 50 features | None | – | 2.4 s | 4 |
VD2 | 50 features | None | – | 408 ms | 2 |
VD1 | 50 features | sc-SynO | 1.23 s | 1.94 s | 5/5 |
VD2 | 50 features | sc-SynO | 1.12 s | 484 ms | 3/3 |
VD1 | 20 features | None | – | 706 ms | 4 |
VD2 | 20 features | None | – | 240 ms | 1 |
VD1 | 20 features | sc-SynO | 1.12 s | 679 ms | 6/5 |
VD2 | 20 features | sc-SynO | 1.11 s | 236 ms | 3/3 |