I had a question about channel attention. I am reading the following paper, that talks about using a seNet.
I am trying to understand the concept of channels. They appear to be referring to more than 3/4 here, ie: more than the standard RGBA channels in an image. Is this in the case of using data that has more than 3/4 channels to represent an image?
They also mention this paper:
Which also refers to multiple channels.
Am I missing something, or does the term ‘channels’ refer to something else other than standard CIE, RGB/RGBA , YUV, HSL/HSV, and CMYK?