-3

I could find max pooling is the most used and preferred type when it comes to Pooling, whatever the image data or the features i need to extract which is sound so ridicules to me for example i'm working on detecting the Diabetic Retinopathy and i need to extract some micro features from the image of retina so why not choosing an average pooling or minimum pooling

moheb yossif
  • 45
  • 7
  • 2
    Minimum pooling doesn't make sense, unless you are looking for a particular feature to be absent, which would be the same as a trained feature to just not activate, so it wouldn't be necessary. Are you analyzing image data or some other form? – Chris Farr Apr 15 '19 at 20:48

1 Answers1

1

Max pooling works better for image data but as you guessed it doesn't necessarily work better for audio data as well. Average pooling is a reasonable alternative for audio data. However it is usually better to try out both and see which one works better for your case. Minimum pooling doesn't make much sense because in neural network we give more weight to features with higher activation.

xashru
  • 3,400
  • 2
  • 17
  • 30