Dout regarding the what happen to 3rd layer of input

#1

sir I have a dout regarding the what happen to 3rd layer of input if we pass (224,224,3) shape array to our sequential CNN model and wehave declared the filter shape=(3,3).
Does shape of filter shape transform automatically to (3,3,3) or what?

#2

hey @tanujvats0002_111433f919bfdb0c,
actually that 3x3 filter is applied one by one on the three channels individually.
R G and then on B.

#3

if the 3*3 filter applied on each channel then the output must be of shape (None,222,222,3,32) then why it is written there that output matrix has dim(None,222,222,32)

#4

each channel doesn’t means the image remains to be in that dimension .
it performs initially on individual channels, but then convert them as being being a single value for that group of pixels