I am somewhat new to image preprocessing and CNNs and was wondering why this is the case. I was testing several random gray images where I believed there did not exist a pure black or pure white portion of the image, though I still got 0 as the min pixel and 255 as the max pixel.
It seems that all images have 0 as the min and 255 as the max even if the image does not display a pure white or black in a grayscale setting.
I tested by the PIL library like this:
x = numpy.asarray(PIL.Image.open('gray.webp')) print(x.min()) print(x.max())
Is pil manipulating the values so that it fits into a [0,255] range, or is something else going on behind the scenes that I do not understand.