3

I am clear with the tasks of instance segmentation and semantic segmentation. However, from the perspective of the neural networks, what is the relationship between them? Namely, is it feasible to realize instance segmentation by improving or modifying a neural network for semantic segmentation, e.g. DeepLab? If so, what operations are usually used? Many thanks.

JJTT
  • 123
  • 1
  • 8

1 Answers1

2

Lets assume that you want to know where exist desired class in image then you makes nn that for each pixel predict probability where is that class - this is sematic segmentation. And when you want to know where exist each instance for desired class this is instance segmentation.picture for example

Dima Komarovski
  • 41
  • 5
  • Well, I know exactly the literal difference, however, I would like to know the differences from the neural network layers of the semantic segmentation and the instance segmentation. Something like that, we have a backbone to achieve semantic segmentation, followed by a classifier for each instance in each class? – JJTT Nov 14 '19 at 13:05
  • There are exist a several ways to make a instance segmentation the most common its a mask-rcnn architecture its works like two step detection(rcnn, fast-rcnn, etc.) but for each anchor also predict segmentation map. – Dima Komarovski Nov 15 '19 at 10:10