I would like to use the bottleneck layer of U-Net (last layer of the encoder) to calculate the similarity between two images. For that I have to somehow flatten the last layer of the encoder. In my opinion there are two approaches:
- Take the last layer which in my case is 4 x 4 x 16 and flatten it to 1D
- Add a dense before the decoder and then reshape the dense 1D layer into 3D
For the second case I am not to sure, how this would affect the network. Arbitrarily reshaping a 1D array into a 3D tensor. Could that introduce weird artifacts? Does someone have experience in a similar problem?