Table 2.

Test error (%) on the MNIST classification task.

Method	Reconstruction	Parameters [K]	MNIST [%]
Our Baseline	No	173	0.48
Base-CapsNet	No	183	0.54
Our Baseline	Yes	173	0.4
Base-CapsNet	Yes	183	0.39
Efficient-CapsNet	Yes	161	$0.{26}_{\pm 0.0002}$
Baseline10	No	35400	0.39
CapsNet10	Yes	6800	$0.{25}_{\pm 0.005}$ ($0.{36}_{\pm 0.04}$)*
Matrix-CapsNet with EM routing14	No	310	0.44
DA-CapsNet29	Yes	7000	0.47
AR CapsNet27	Yes	5310	0.54
HFCs20	No	1514	$0.{25}_{\pm 0.0002}$

All methodologies are reported with their number of parameters and the presence of the reconstruction regularizer during the training phase. * indicates the results from our experiments.