Skip to main content
Cross Validated

Return to Answer

added 18 characters in body
Source Link
doubllle
doubllle
  • 1.9k
  • 1
  • 16
  • 21

My two cents on understanding the 1x1 convolution:

It is a two dimension version of dense layer

From 192x28x28 to 64x28x28, it is like a dense layer with 192 inputs and 64 outputs, and this layer is applied to 28x28 image. So the calculation is a linear or nonlinearperforming feature maps combination and dimension reduction on the axis of 192.

My two cents on understanding the 1x1 convolution:

It is a two dimension version of dense layer

From 192x28x28 to 64x28x28, it is like a dense layer with 192 inputs and 64 outputs, and this layer is applied to 28x28 image. So the calculation is a linear or nonlinear dimension reduction on the axis of 192.

My two cents on understanding the 1x1 convolution:

It is a two dimension version of dense layer

From 192x28x28 to 64x28x28, it is like a dense layer with 192 inputs and 64 outputs, and this layer is applied to 28x28 image. So the calculation is performing feature maps combination and dimension reduction on the axis of 192.

Source Link
doubllle
doubllle
  • 1.9k
  • 1
  • 16
  • 21

My two cents on understanding the 1x1 convolution:

It is a two dimension version of dense layer

From 192x28x28 to 64x28x28, it is like a dense layer with 192 inputs and 64 outputs, and this layer is applied to 28x28 image. So the calculation is a linear or nonlinear dimension reduction on the axis of 192.