2
$\begingroup$

I'm studying for a computer vision module and I'm on the deep learning topic, in one past paper we have the following question:

Given that a convolutional neural network has five convolution layers (all the convolution layers are composed of 3×3 convolution filters with stride 1 and no pooling layers), calculate how many pixel(s) in the input image are supporting a neuron in the fifth layer?

without the input size of the image, how is it possible to work this out?

Improve this question
$\endgroup$
1

1 Answer 1

Reset to default
2
$\begingroup$

If there is only one convolution layer, then every neuron is supported by $3 \times 3 = 9$ pixels in the input image. This doesn't depend on the size of the image. You can generalize this to more layers.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Ok, that makes sense thank you! So we would take the number of layers * the number of pixels supported by the first layer, given all the layers in the network are convolutional layers? is this correct or would we do the number of pixels to the power of the number of layers? $\endgroup$
    – Yallooss
    Commented Jun 2, 2021 at 12:34
  • 2
    $\begingroup$ I think you now have enough information to solve the problem on your own. $\endgroup$
    – Yuval Filmus
    Commented Jun 2, 2021 at 12:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.