TABLE I.

Summary of the patch sizes for all the different CNNs utilized in this work. Ch. stands for channels, MI for multi-inputs (indicating that CT and enhanced images are used as separate inputs to the network), and # is used for number.

	# Ch.	Enhanced images (Y/N)	# of patches	# patches × (x, y, depth, ch.)
Proposed
3D CT only	1	N	1	1×(32,32,5,1)
Alternatives
1. 2.5D CT only	1	N	1	1×(32,32,1,3)
2. 2.5D 3 Ch.	9	Y	1	1×(32,32,1,9)
3. 2.5D MI	3	Y	3	3×(32,32,1,3)
4. 3D 3 Ch.	3	Y	1	1×(32,32,5,3)
5. 3D MI	1	Y	3	3×(32,32,5,1)