In Implementation of UNet Network the input dimensions are (4, 160, 160, 16) ,which are number of channels , height , width and length . What does the length dimension signify ?
This is a three dimensional input. Here, 16 in the depth or z of the three dimensional structure.
Don’t you think 4 is the depth where 16 is the number of samples
@sbansal793 when we take input with any frameworks or tools, they take first input as kernel*width*height*depth. And it hardly happens that depth is lower than the number of kernels.
Do you mean by kernels is number of samples?
Yes. You can call feature too.