Is the normalization of an image in double a lossless operation?

Question

Lets say I have an image in double that goes from range 0.0 to 32.0. If I normalize it to get the range 0.0 to 1.0, by dividing by the max value from the image, do I lose information? I don't know if it changes anything but I do this in c++.

For the format at first talk about jpeg, but then after if we talk about the signal itself, if I were to lets say apply a k-mean to that "array" of double, would it make a difference if it were normalized?

Answer 1

缩放浮点数不会影响超过1个最低有效位，这在图像处理的上下文中没有影响。