I’m guessing that you are thinking of convolution, not pooling. Pooling works differently.
Welcome to our Community! Thanks for reaching out. We are here to help you.
Additional to @bs80 response. The pooling operation reduces the spatial dimensions of the input image by a factor of the pool size. In this case, the pool size is 2x2, so the resulting image will have half the height and half the width of the input image.