receptive field calculation after the deconvolution layers

[ntajbakhsh]

Hi Fausto,

It was pretty straight forward for me to calculate the receptive field for L-Stage 1,2,3,4,5, but I'm having difficulty working out the receptive field of R-Stage 4 as reported in Table 1 of your paper. I guess I'm not quite sure how the the deconvolution layers and the maps forwarded from the left part of the net contribute to the increased receptive field. Could you provide some hints or point me to a reference that does such calculations?

Many thanks, Nima

[sagarhukkire]

@ntajbakhsh

Check closed issue where I mentioned the paper for this solution It is blockwise calculation but in paper you will get perfect idea

[ntajbakhsh]

@sagarhukkire Thanks for your reply. I read the closed thread but didn't find a link to the paper. Could you please point me to the reference?

[sagarhukkire]

@ntajbakhsh

https://arxiv.org/pdf/1701.03056.pdf

[ntajbakhsh]

Thanks!

On Wed, May 3, 2017 at 2:41 PM, Sagar Hukkire notifications@github.com wrote:

@ntajbakhsh https://github.com/ntajbakhsh

https://arxiv.org/pdf/1701.03056.pdf

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/faustomilletari/VNet/issues/36#issuecomment-299044077, or mute the thread https://github.com/notifications/unsubscribe-auth/AMD9qV7kYIRrJ9YM5vmRA5iEwWSbi21Aks5r2PSjgaJpZM4NNa5o .

--

Nima Tajbakhsh, Ph.D Email: Nima@asu.edu Nima@asu.edu Webpage: http://www.public.asu.edu/~ntajbakh/ http://www.public.asu.edu/~ntajbakh/Cell: +1 (480) 512 1286

[rogertrullo]

Hi @ntajbakhsh, I am also trying to get the receptive field. So far I've only been able to do it in the down-sampling path, but no luck with the up-sampling (deconvolutions). Were you able to compute it?

[caney1989]

Hi! @ntajbakhsh , has your problem been solved? I have been troubled for how to caculate the receptive field for the deconvolution layers. I have caculated the receptive field for the networks designd in V-Net (http://arxiv.org/abs/1606.04797) and CNN-based Segmentation of Medical Imaging Data (http://arxiv.org/abs/1701.03056) , but i can not find the right way to caculate the receptive field for the deconvolution layers. Could you show me some details for this caculation? Thanks very much!!