wsss_sam

PyTorch implementation of An Alternative to WSSS? An Empirical Study of the Segment Anything Model (SAM) on Weakly-Supervised Semantic Segmentation Problems

This is an inference-only method, no training is needed.

Install

You should set the environment variable manually as follows if you want to build a local GPU environment for Grounded-SAM:

export AM_I_DOCKER=False
export BUILD_WITH_CUDA=True
export CUDA_HOME=/path/to/cuda-11.3/

Install Segment Anything:

python -m pip install -e segment_anything

Install Grounding DINO:

python -m pip install -e GroundingDINO

Download the pretrained weights

cd wsss_sam

wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth
wget https://github.com/IDEA-Research/GroundingDINO/releases/download/v0.1.0-alpha/groundingdino_swint_ogc.pth

Running

Prepare pascal and COCO data.

Set the data path accordingly and run:

bash run_coco.sh

bash run_pascal.sh

Citation

@misc{sun2023alternative,
      title={An Alternative to WSSS? An Empirical Study of the Segment Anything Model (SAM) on Weakly-Supervised Semantic Segmentation Problems}, 
      author={Weixuan Sun and Zheyuan Liu and Yanhao Zhang and Yiran Zhong and Nick Barnes},
      year={2023},
      eprint={2305.01586},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Acknowledgement

This work is heavily built upon the codebase provided by Grounded-Segment-Anything, Thanks for their great code.