Open Access

Table 2

Number of mock galaxies, real sources, and false positives detected by each source-finding method prior to applying the random forest classifier.

Frequency range (MHz) Mock galaxies Real sources False positives
SoFiA MTO V-Net SoFiA MTO V-Net SoFiA MTO V-Net
Training 1280–1304 0 (0%) 2 (1%) 40 (13%) 0 1 1 33 2069 44
Pointing 1300–1324 15 (5%) 1 (0%) 93 (31%) 10 0 11 74 0 34
1320–1344 54 (18%) 6 (2%) 145 (48%) 3 0 2 52 1 35
1341–1364 85 (28%) 20 (7%) 199 (66%) 15 3 14 32 4 25
1361–1385 175 (58%) 66 (22%) 258 (86%) 10 3 8 156 10 13
1381–1405 268 (89%) 190 (63%) 292 (97%) 41 21 35 70 9 12
Total 597 285 1027 79 28 71 417 2093 163
Test 1280–1304 1 (0%) 1 (0%) 7 (2%) 0 0 0 23 3094 1030
Pointing 1300–1324 23 (8%) 0 (0%) 34 (11%) 23 1 11 184 13 20
1320–1344 61 (20%) 5 (2%) 55 (18%) 1 0 0 65 12 21
1341–1364 108 (36%) 26 (9%) 111 (37%) 40 4 20 79 65 40
1361–1385 154 (51%) 79 (26%) 193 (64%) 24 7 17 13 64 33
1381–1405 279 (93%) 199 (66%) 278 (93%) 24 14 18 63 240 29
Total 626 310 678 112 26 66 427 3488 1173

Notes. These detections are shown for each spectral window of the Apertif training and test pointings. For the mock galaxies, the percentage of the total number of mock galaxies in the spectral window, 300, is shown in brackets.

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.