A deep learning approach for automated segmentation of magnetic bright points in the solar photosphere

Home

All issues

Volume 677 (September 2023)

A&A, 677 (2023) A121

Full HTML

Open Access

Table 4

Testing metrics of our method and that of Yang et al. (2019).

	Our method					Yang et al. (2019)

Data time	Precision	Recall	F1	PA	IoU	Precision	Recall	F1	PA	IoU
2018-06-07 22:04:00 UT	1.0	0.817	0.899	0.998	0.817	0.823	0.449	0.581	0.992	0.410

2019-06-18 16:37:34 UT	0.984	0.800	0.880	0.999	0.785	0.931	0.387	0.547	0.996	0.376

2019-06-24 20:30:05 UT	0.970	0.731	0.834	0.998	0.715	0.935	0.238	0.380	0.995	0.234

2020-07-18 16:58:30 UT	0.989	0.851	0.915	0.999	0.843	0.949	0.373	0.536	0.997	0.366

2021-04-30 18:29:18 UT	0.936	0.937	0.936	0.999	0.880	0.940	0.424	0.584	0.997	0.413

mean	0.976	0.827	0.893	0.999	0.808	0.916	0.374	0.526	0.995	0.360

Notes. For ease of comparison, we highlight in bold font the average highest indicator accuracy for each time span.

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.