1

Consider the following sentences from the research paper titled PatternNet: Visual Pattern Mining with Deep Neural Network by Hongzhi Li et al.

The value of each pixel in a feature map is the response of a filter with respect to a sub-region of an input image. A high value for a pixel means that the filter is activated by the content of the sub-region of the input image.

The sentences mention the sub-regions of an image. Is there any formal definition for a sub-region of an image?

hanugm
  • 3,571
  • 3
  • 18
  • 50

1 Answers1

1

It seems that they are informally using the term "sub-region" to refer to the section of the image with which you multiply the kernel to produce a scalar value of the feature map (which they call pixel, but I would use the term pixel only to refer to the scalar values of the image). So, it seems that a sub-region is a synonym for receptive field. We have a question on the difference between receptive fields and feature maps here.

nbro
  • 39,006
  • 12
  • 98
  • 176