About how you preprocess the image

[breadbread1984]

I saw you preprocess input image by subtracting the mean intensity of it rather than subtracting the mean intensity of the whole training images as people usually does. What is the difference? Any reason for it?

[luoyetx]

It gives a little better performance than pixel / 256 or (pixel - 128) / 128 when I train the network. There's no particular reason I choose this method. If I need to train the model again, I will choose a simple processing (pixel - 128) / 128. I was too naive that time that only focused on the data processing not the optimization itself. I think the training itself matters more.