YOLO-CIANNA: Galaxy detection with deep learning in radio data - II. Winning the SKA SDC2 using a generalized 3D-YOLO network

Open Access

Table C.1

Detection-only score metrics with and without the extra-parameter prediction enabled during training.

Model	AP_−0.3	AP_0.1	AP_0.5	$N_{- 0.3}^{m}$ $Mathematical equation: $\[N_{-0.3}^{m}\]$$	$N_{0.1}^{m}$ $Mathematical equation: $\[N_{0.1}^{m}\]$$	$N_{0.5}^{m}$ $Mathematical equation: $\[N_{0.5}^{m}\]$$	AP_SDC2	$N_{SDC2}^{m}$ $Mathematical equation: $\[N_{\text {SDC2}}^{m}\]$$	$\bar{s}$ $Mathematical equation: $\[\bar{s}\]$$
With param. BASE	19.20	16.72	9.01	54 022	46 040	27 476	18.35	51 520	0.7816
With param. BT1	19.34	16.99	9.70	52 524	45 323	28 567	18.32	49 306	0.7908
With param. BT2	19.33	16.87	9.29	52 768	45 302	27 806	18.34	49 579	0.7903
With param. BT3	19.26	16.89	9.64	52 672	45 282	28 407	18.25	49 437	0.7910
No-param.	19.24	16.69	8.18	55 996	46 997	26 807	N/A	N/A	N/A

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.