TABLE 1.

Values referring to the images used for datasets creation, both synthetic and real, for segmentation and rotation Convolutional Neural Networks

			Segmentation	Rotation
Surgical operation	Case study	Target	# Real images	# Synthetic images	Rotation ranges in degrees (−X,+X)/(−Y,+Y)/(−Z, +Z)
RARP	1	Catheter	375	40000	(−40, 10)/‐/‐
RARP	2	Prostate	388	35000	(−15, 20)/(−25, 25)/(−5, 15)
RAPN	3	Kidney	208	40000	(−10, 10)/(−10, 10)/(−10, 10)

Note: For the latter, considered rotation ranges are also shown. For the catheter, the Rotation CNN was trained to predict only the X rotation, as Y was derived directly from the segmentation map and Z was considered irrelevant.