Question

Consider a CNN that takes in 32 × 32 grayscale images and has a single convolution layer with three 5 × 5 convolution filters (without boundary padding). What is the output size of the convolutional layer?

a) 30 × 30
b) 28 × 28
c) 26 × 26
d) 25 × 25

Answer 1

The output size of a convolutional layer with a 5 × 5 filter applied to a 32 × 32 grayscale image without padding is 28 × 28.

Step-by-step explanation:

The student is asking about the output size of a convolutional layer in a Convolutional Neural Network (CNN). Given that a single convolution layer with three 5 × 5 convolution filters is applied to a 32 × 32 grayscale image without any boundary padding, the output size can be determined by the formula:

(W - F + 1) × (H - F + 1)

where W is the width of the image (32), H is the height of the image (32), and F is the filter size (5). Substituting the values gives:

(32 - 5 + 1) × (32 - 5 + 1) = (28) × (28)

Therefore, the output size of the convolutional layer is b) 28 × 28.