search

Unrealluver / dgcn

Code for paper "Deep Graph Cut Network for Weakly-supervised Semantic Segmentation" (http://xinggangw.info/pubs/scis-dgcn.pdf)
3 stars 0 forks source link

how to generate cues? #2

Open tangmw opened 3 years ago

tangmw commented 3 years ago

We want to try this project with different CAM methods, but we dont know how to generate cues（localization_cues-sal.pickle）. Can you release this part code? Thanks

Unrealluver commented 3 years ago

Thank you for your interest for our repo! About the method to generate the cues, I have add it in the newest commit. You may find it in the Readme.md at the Cues' part!

tangmw commented 3 years ago

Thanks for your reply. But we still dont know how to blend the foreground and background regions to gain cues. Just concatenate them?