Euclid preparation - XXXIII. Characterization of convolutional neural networks for the identification of galaxy-galaxy strong-lensing events

Open Access

Table B.3

Summary of the performance of the VGG-like network, the inception network, and the residual network, trained on the selection S2, in classifying the objects that are part of the selections S3 and S4.

	VGG-like network				Inception Network				Residual Network

	S2/S3		S2/S4		S2/S3		S2/S4		S2/S3		S2/S4
Class	0	1	0	1	0	1	0	1	0	1	0	1

Precision	0.77	0.97	0.62	0.95	0.82	0.96	0.65	0.93	0.75	0.88	0.64	0.85

Recall	0.98	0.68	0.98	0.33	0.97	0.76	0.97	0.42	0.92	0.67	0.94	0.40

F1-score	0.86	0.80	0.76	0.48	0.89	0.85	0.78	0.58	0.83	0.76	0.76	0.55

Accuracy	0.83		0.68		0.87		0.71		0.80		0.69

AUC	0.57		0.52		0.81		0.7		0.78		0.65

Notes. The precision, recall, and F1-score are evaluated on the class of the nonlenses (0) and of the lenses (1) separately.

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.