שאלה | In the context of facial recognition, why is a Siamese network architecture considered advantageous over a "regular" network architecture?

In the context of facial recognition, why is a Siamese network architecture considered advantageous over a "regular" network architecture?

		Siamese networks require fewer computational resources.
		Siamese networks predict the probability of each class for a given picture.
		Siamese networks can compare similarities between pairs of photos and identify individuals not trained on.
		"Regular" networks are more effective in recognizing faces.