0
$\begingroup$

What is the Number of capsules in the Primary Capsule Layer of Capsule networks?

In many articles, it is written that the number of Capsules is 32 but in the paper, by Hinton - Dynamic Routing between capsules it is written that "In total PrimaryCapsules has [32 × 6 × 6] capsule outputs (each output is an 8D vector) and each capsule in the [6 × 6] grid is sharing their weights with each other."

which implies that the layer has 36 capsules.

$\endgroup$

1 Answer 1

0
$\begingroup$

It could be that each capsule is of 2D form (thus 6x6). So each component capsules outputs an 8D vector. The 8D vector is derived by analyzing 8 feature maps propagated from the convolutional layer. So its like placing the feature maps in groups of 8. This means each component capsule represents one capsule therefore in the "dynamic paper", there are actually 32 capsules of form 8D.

NB: The Primary Capsule Layer is actually made up of convolutional layers. So you can look at the 6x6 as the size of the feature maps after convolution in the primary Capsule layer.

take a look at this site. Anyway I am still reading wide on capsules, because most of the stuffs in the paper are consolidated. I am open to corrections as well. Thanks

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.