search

YudeWang / SEAM

Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentation, CVPR 2020 (Oral)
MIT License
542 stars 97 forks source link

Performance from the provided weights #10

Closed halbielee closed 4 years ago

halbielee commented 4 years ago

Hello!

Thanks for the sharing the code.

I ran your code with the trained weights and got lower performance than the paper reported. 60.076 % mIOU for the validation set.

My inference step is 1) Infer the CAM(npy files) from the infer_SEAM.py with the trained model 2) Infer the Segmentation map(png files) from the infer_aff.py with the trained model 3) evaluate the png files with gt files.

Is there anything I missed? Than you!

YudeWang commented 4 years ago

Hi @halbielee , The generated pseudo segmentation labels are evaluated on PASCAL VOC 2012 train set (pixel-level gt of train set are not available during training). The segmentation model is retrained on these pseudo segmentation labels then evaluated on both validation and test set.

halbielee commented 4 years ago

Oh! You are right.

I was confused that the result of Affinitynet was a segmentation result.

Thanks for the reply!