Table 1.

Number of trainable parameters per layer in each model.

Layer	Iterative method with 3-D convolutions	Iterative method with 2-D convolutions	2-D method	3-D method
1	896	1152	384	896
2	55,488	55,872	18,624	55,488
3	221,568	222,336	74,112	221,568
4	885,504	887,040	295,680	885,504
5	295,296	885,120	295,296	885,120
6	442,752	1,032,576	442,752	1,327,488
7	147,648	295,104	147,648	442,550
8	36,960	73,824	36,960	110,688
9	9264	18,480	9264	13,872
10	588	588	588	1740
11	78	78	78	222
12	19	19	19	55
Total	2,096,061	3,472,189	1,321,405	3,945,191