Zero Padding for Neural Networks - Deep Learning Dictionary
Zero Padding - Deep Learning Dictionary
Convolutional layers in a CNN have a defined number of filters that convolve the image input, and as a result, the output from the convolutional layers have reduced dimensions from the original image input.
We have a general formula for determining the dimensions of the convolved output. Assuming a stride of \(1\), if our image is of size \(n \times n\), and we convolve it with an \(f \times f\) filter, then the dimensions of the resulting output is
Committed by on