How to create 1d image input layer to use convolution1dLayer?

I am trying to use convolution1dLayer on 1D image inputs. The documentation of convolution1dLayer says

• For 1-D image input (data with three dimensions corresponding to the spatial pixels, channels, and observations), the layer convolves over the spatial dimension.

So its input format is "SCB". However, I didn't find an inputlayer that output data in the "SCB" format. The imageInputLayer can only output data in "SSCB". How should I input my 1D-image data to feed into convolution1dLayer? Many thanks for any helps in advance!

Best,

Xiaohao

Answer 1

[Conor Daly]

When working with dlnetwork input layers are not required. So you can work with "SCB" data by directly injecting into the network at the appropriate layer.

The repo linked below contains a network which operates on "SCB" data:

https://github.com/matlab-deep-learning/fourier-neural-operator

[Xiaohao Sun]

Interesting. Many thanks for your answer and example!

Also, it seems the functionLayer can be used in layer array to reformat the data.