mrgloom / awesome-semantic-segmentation

:metal: awesome-semantic-segmentation

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Why is this awesome?

apacha opened this issue · comments

Firstly, I would like to thank you for your efforts in compiling this list.

I was just wondering, why this is an awesome list for semantic segmentation? It more feels like a collection of resources for semantic segmentation, rather than a curation. It lacks a critical review and qualified judgement of the referenced articles and tools.

As someone who would like to start with Semantic Segmentation, I would be totally lost in this pile of links:

  • Where would you start?
  • What is the current state-of-the-art?
  • What has been shown to work well / not work well? Ideally you would start with a recent review paper over the different methodologies. Is there such a paper?
  • It would also be great to put publication dates next to papers or sort them chronologically to see what is well-established and what is still bleeding edge.

This list is more about towards practice and reproducible research then theoretical top state of the art results.
Actually I'm planning to create one more benchmark like already done here
https://github.com/mrgloom/awesome-semantic-segmentation#benchmarks

As someone who would like to start with Semantic Segmentation, I would be totally lost in this pile of links

I recommend to start with network architectures that is more popular(have more implementations), like Unet for example.

What is the current state-of-the-art

This section should answer the question
https://github.com/mrgloom/awesome-semantic-segmentation#results

Ideally you would start with a recent review paper over the different methodologies

Maybe I will add section with survey papers.

It would also be great to put publication dates next to papers or sort them chronologically to see what is well-established and what is still bleeding edge.

Yes, it's a good idea.