cvlab-yonsei / BANA

An official implementation of "Background-Aware Pooling and Noise-Aware Loss for Weakly-Supervised Semantic Segmentation" (CVPR 2021) in PyTorch.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

BANA

This is the implementation of the paper "Background-Aware Pooling and Noise-Aware Loss for Weakly-Supervised Semantic Segmentation".

For more information, please checkout the project site [website] and the paper [arXiv].

Requirements

Getting started

The folder data should be like this

    data   
    └── VOCdevkit
        └── VOC2012
            ├── JPEGImages
            ├── SegmentationClassAug
            ├── Annotations
            ├── ImageSets
            ├── BgMaskfromBoxes
            └── Generation
                ├── Ycrf
                └── Yret
git clone https://github.com/cvlab-yonsei/BANA.git
cd BANA
python stage1.py --config-file configs/stage1.yml --gpu-id 0 # For training a classification network
python stage2.py --config-file configs/stage2.yml --gpu-id 0 # For generating pseudo labels
python stage3_vgg.py --config-file configs/stage3_vgg.yml --gpu-id 0 # For training DeepLab-LargeFOV

Download our pseudo labels and BgMaskfromBoxes on PASCAL VOC 2012

Bibtex

@inproceedings{oh2021background,
  title     = {Background-Aware Pooling and Noise-Aware Loss for Weakly-Supervised Semantic Segmentation},
  author    = {Oh, Youngmin and Kim, Beomjun and Ham, Bumsub},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2021},
}

About

An official implementation of "Background-Aware Pooling and Noise-Aware Loss for Weakly-Supervised Semantic Segmentation" (CVPR 2021) in PyTorch.

License:GNU General Public License v3.0


Languages

Language:Python 100.0%