Bumped by Community user
I reformulate the title to make it more compelling.
Link
nbro
  • 31.2k
  • 8
  • 65
  • 130

What value How should we pad an image to addbe fed in border areaa CNN?

added 81 characters in body
Source Link

As everyone experienced in deep learning might know, in an image classification problem we normally add borders to images before feed them intothen resize it to the input size of a CNN network in order. The reason of doing this is to keep aspect ratio of the original image and retain it's information.

I have seen people fill black (0 pixel value for each channel), gray (127 pixel value for each channel), or random value generated from gaussian distribution to the border.

My question is, is there any proof that which of these is correct?

As everyone experienced in deep learning might know, in an image classification problem we normally add borders to images before feed them into the network in order to keep aspect ratio.

I have seen people fill black (0 pixel value for each channel), gray (127 pixel value for each channel), or random value generated from gaussian distribution to the border.

My question is, is there any proof that which of these is correct?

As everyone experienced in deep learning might know, in an image classification problem we normally add borders to images then resize it to the input size of a CNN network. The reason of doing this is to keep aspect ratio of the original image and retain it's information.

I have seen people fill black (0 pixel value for each channel), gray (127 pixel value for each channel), or random value generated from gaussian distribution to the border.

My question is, is there any proof that which of these is correct?

Source Link

What value to add in border area?

As everyone experienced in deep learning might know, in an image classification problem we normally add borders to images before feed them into the network in order to keep aspect ratio.

I have seen people fill black (0 pixel value for each channel), gray (127 pixel value for each channel), or random value generated from gaussian distribution to the border.

My question is, is there any proof that which of these is correct?