search

lingorX / HieraSeg

CVPR2022 - Deep Hierarchical Semantic Segmentation - A structured, pixel-wise description of visual scenes in terms of the class hierarchy.
262 stars 24 forks source link

Loss Problem #9

Open YuSha-FIAS opened 9 months ago

YuSha-FIAS commented 9 months ago

Hello, this is very nice work. But I have some problems.

  1. /mmseg/models/losses/cross_entropy_loss.py is equal L_BCE;
  2. /mmseg/models/losses/hiera_loss.py include losses_hiera and losses_hiera_focal. Losses_hiera is L_TM, right? And losses_hiera_focal is L_FTM, right?
  3. /mmseg/models/losses/tree_triplet_loss.py is L_TT, right?
  4. /mmseg/models/losses/rmi_hiera_triplet_loss.py, please What is rmi_hiera_triplet_loss?
  5. What is lovasz_loss?
  6. hiera_triplet_loss_cityscape.py for Cityscapes dataset, right?