For performing the code generation for “image3dInputLayer”, you can replace “image3dInputLayer” with “imageInputLayer” and manually reshape or handle the data appropriately.
For example, let’s say we are having the below code snippet using “image3dInputLayer”:
image3dInputLayer([32 32 16 4], 'Name', 'input')
Since “ imageInputLayer” does not support 4D inputs, there is a need to combine depth and channels into a single dimension, thus, depth × channels = 16 × 4 = 64.
imageInputLayer([32 32 64], 'Name', 'input', 'Normalization', 'none')
You can refer to the below MATLAB documentation for code generation of “dlnetwork”:
I hope this helps.
