What is difference between Flatten() and Dense() layers in Convolutional Neural Network?

Question

I have confusion between the two. Could someone please elaborate on it with some examples?

https://stackoverflow.com/questions/43237124/role-of-flatten-in-keras — yaho cho, May 23 '19 at 16:37

score 5 · Answer 1 · answered May 23 '19 at 15:46

Flatten as the name implies, converts your multidimensional matrices (Batch.Size x Img.W x Img.H x Kernel.Size) to a nice single 2-dimensional matrix: (Batch.Size x (Img.W x Img.H x Kernel.Size)). During backpropagation it also converts back your delta of size (Batch.Size x (Img.W x Img.H x Kernel.Size)) to the original (Batch.Size x Img.W x Img.H x Kernel.Size).

Dense layer is of course the standard fully connected layer.

What is difference between Flatten() and Dense() layers in Convolutional Neural Network?

1 Answers1