Open Access

Table 3

Macro average multi-class metrics scores (%) on MACHO, ATLAS, and OGLE-III test sets.

Method Metric MACHO ATLAS OGLE-III
Baseline F1 68.6 ± 1.7 77.8 ± 2.6 67.3 ± 2.9
Accuracy 69.5 ± 1.6 77.7 ± 2.5 68.5 ± 2.9
Precision 69.3 ± 1.7 78.1 ± 2.6 68.6 ± 2.8

MC Dropout F1 68.0 ± 0.6 78.8 ± 0.7 66.8 ± 1.6
Accuracy 69.3 ± 0.5 78.8 ± 0.7 68.4 ± 1.4
Precision 68.4 ± 0.6 79.3 ± 0.8 67.4 ± 1.6

HSA F1 74.6 ± 1.8 82.1 ± 2.2 75.7 ± 1.9
Accuracy 75.3 ± 1.8 82.2 ± 2.1 76.7 ± 1.7
Precision 75.3 ± 1.6 82.9 ± 1.8 77.3 ± 1.3

HA-MC Dropout F1 76.6 ± 0.8 84.0 ± 1.3 79.8 ± 0.5
Accuracy 77.5 ± 0.7 84.0 ± 1.3 80.5 ± 0.4
Precision 77.1 ± 0.5 84.3 ± 1.2 80.0 ± 0.3

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.