Closed sooyeonshin closed 7 years ago
Hi I have bin interest about your method.
can i ask you one question?
I don't under stand about last layer.
last layer got 64 filtered image at previous layer and use 64 filters for convolution.
but how reconstruct one image?
reconstruct image pixel is sum of each 64 filter image pixel?
The filter size is 3X3X64X1. This is the basic knowledge of CNN. See http://www.robots.ox.ac.uk/~vgg/practicals/cnn-reg/index.html for more details.
Hi I have bin interest about your method.
can i ask you one question?
I don't under stand about last layer.
last layer got 64 filtered image at previous layer and use 64 filters for convolution.
but how reconstruct one image?
reconstruct image pixel is sum of each 64 filter image pixel?