Convert rgb image vector into grey scale image

I read the rgb image in tensor lets say “rgb_image” with dimension [1,3,256,256]. How can i convert it into the grey scale image?

Hi cbd!

The most standard way:

grayscale_image = rgb_image.mean (dim = 1)

grayscale_image will now be of shape torch.Size([1, 256, 256]).

The idea is that “3” dimension of rgb_image consists of the three
color channels, r, g, and b. You take their mean (or sum them, if
you prefer) to get the “brightness” of the pixel in question, that is,
its grayscale level.


K. Frank