ACR_WSSS

[ICCV 2023 Workshop] Official implementation of the paper: All-pairs Consistency Learning for Weakly Supervised Semantic Segmentation

We won the best paper award of the 1st Workshop on New Ideas in Vision Transformers at ICCV 2023!

Step1: environment

• clone this repo

git clone https://github.com/OpenNLPLab/ACR_WSSS.git

• optionally create a new environment python>=3.6
• install requirements.txt

pip install -r requirements.txt

Step2: dataset preparation

pascal voc

MS-COCO 2014

Step3: train and inference

First, change your data path and GPU settings accordingly. Then you can perform train and inference Run:

bash train_acr.sh

which includes train, localization generation and evaluation.

Acknowledgement

• Thanks for codebase provided by DPT
• Thanks for codebase provided by RRM

---

if you use this paper, please kindly cite:

@misc{sun2023allpairs,
      title={All-pairs Consistency Learning for Weakly Supervised Semantic Segmentation}, 
      author={Weixuan Sun and Yanhao Zhang and Zhen Qin and Zheyuan Liu and Lin Cheng and Fanyi Wang and Yiran Zhong and Nick Barnes},
      year={2023},
      eprint={2308.04321},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}