Where do the numbers come from in the net surgery example?

[Peter Wolf]

In the net surgery example, the text says:

>The fc6 inner product when carried out as convolution by fc6-conv turns into a 6 \times 6 filter with stride 1 on pool5. Back in image space this gives a classification for each 227 × 227 box with stride 32 in pixels.

I see that ((451 - 227) / 32) + 1 = 8, but where exactly do the 227 and/or the 32 come from? The 451 is defined about the input-image size and the 8 is given by the dimension of the fc6-conv layer.
Is the "kernel" 227 because that was the image size the net was trained on?

A short explanation or link to one would be appreciated.

[Peter Wolf]

To answer my own question: The stride of 32 results from multiplying all strides of the convolutional and pool layers.
4 * 2 * 1 * 2 * 2 = 32

The 227x227 by calculating what a 1x1 area out of the 8x8 feature vector represents in the original image. Just use the formula "backwards".