search

xmed-lab / URN

AAAI 2022: Uncertainty Estimation via Response Scaling for Pseudo-Mask Noise Mitigation in Weakly-Supervised Semantic Segmentation
Apache License 2.0
38 stars 4 forks source link

The generation of published ''Pseduo-masks from classification phase'' #4

Closed Zhengyang1995 closed 2 years ago

Zhengyang1995 commented 2 years ago

Hey, thank you for your impressive works. Would you please kindly give us an explanation to the published ''Pseduo-masks from classification phase''?

I guess it is the CAM generated from the trained classification network, and then post-processed by dCRF, finally with a threshold to distinguish background and certain class. If the procedure is correct, would you please let me know the threshold between bg and fg? Or if there is anything different, please let me know.

Thank you so much!

Zhengyang1995 commented 2 years ago

Or maybe there is no dCRF in that moment?

Eli-YiLi commented 2 years ago

Hey, thank you for your impressive works. Would you please kindly give us an explanation to the published ''Pseduo-masks from classification phase''?

I guess it is the CAM generated from the trained classification network, and then post-processed by dCRF, finally with a threshold to distinguish background and certain class. If the procedure is correct, would you please let me know the threshold between bg and fg? Or if there is anything different, please let me know.

Thank you so much!

Your understanding is right, it comes from classification net with post-processing of PPMG with CRF.