Open Access

Table 5.

Performance of the CNNs averaged over the five test sets containing visually classified galaxies from the HSC-SSP survey.

Precision Recall
CNN model z-bin 1 z-bin 2 z-bin 3 z-bin 1 z-bin 2 z-bin 3
[0.1; 0.31) [0.31; 0.52) [0.52; 0.76) [0.1; 0.31) [0.31; 0.52) [0.52; 0.76)
TNG-CNN 0.78 ± 0.02 0.67 ± 0.02 0.61 ± 0.02 0.74 ± 0.01 0.67 ± 0.02 0.81 ± 0.01
Horizon-CNN 0.71 ± 0.01 0.76 ± 0.02 0.82 ± 0.01 0.84 ± 0.01 0.76 ± 0.01 0.63 ± 0.01
Comb-CNN 0.86 ± 0.02 0.80 ± 0.03 0.85 ± 0.02 0.68 ± 0.01 0.58 ± 0.02 0.56 ± 0.01
Comb-CNN-2thr. 0.86±0.01 0.80±0.03 0.85±0.02 0.88±0.01 0.81±0.03 0.80±0.03

F1-score Accuracy

TNG-CNN 0.76 ± 0.01 0.67 ± 0.02 0.69 ± 0.01 0.76 ± 0.01 0.67 ± 0.02 0.64 ± 0.02
Horizon-CNN 0.77 ± 0.01 0.76 ± 0.01 0.71 ± 0.01 0.75 ± 0.01 0.76 ± 0.02 0.75 ± 0.01
Comb-CNN 0.76 ± 0.01 0.67 ± 0.02 0.68 ± 0.01 0.79 ± 0.01 0.72 ± 0.02 0.73 ± 0.01
Comb-CNN-2thr. 0.87±0.01 0.80±0.02 0.82±0.03 0.86±0.01 0.77±0.02 0.82±0.03

Notes. The reported errors correspond to the standard deviation. Precision, recall, and F1-score refer only to the merger class.

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.