Foreground-Covering Prototype Generation and Matching for SAM-Aided Few-Shot Segmentation (FCP)

This is the official repository for the following paper:

Foreground-Covering Prototype Generation and Matching for SAM-Aided Few-Shot Segmentation [Arxiv]

Suho Park*, SuBeen Lee*, Hyun Seok Seong, Jaejoon Yoo, Jae-Pil Heo (*: equal contribution)
Accepted by AAAI 2025

Requirements

• Python 3.10
• PyTorch 1.12
• cuda 11.6

Conda environment settings:

conda create -n fcp python=3.10
conda activate fcp

conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.6 -c pytorch -c conda-forge

Segment-Anything-Model setting:

cd ./segment-anything
pip install -v -e .
cd ..

Preparing Few-Shot Segmentation Datasets

Download following datasets:

1. PASCAL-5ⁱ

Download PASCAL VOC2012 devkit (train/val data):

wget http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar

Download PASCAL VOC2012 SDS extended mask annotations from our [Google Drive].

2. COCO-20ⁱ

Download COCO2014 train/val images and annotations:

wget http://images.cocodataset.org/zips/train2014.zip
wget http://images.cocodataset.org/zips/val2014.zip
wget http://images.cocodataset.org/annotations/annotations_trainval2014.zip

Download COCO2014 train/val annotations from our Google Drive: [train2014.zip], [val2014.zip]. (and locate both train2014/ and val2014/ under annotations/ directory).

3. Image Encoder weights

Resnet : https://drive.google.com/drive/folders/1Hrz1wOxOZm4nIIS7UMJeL79AQrdvpj6v
VGG : https://download.pytorch.org/models/vgg16_bn-6c64b313.pth

Create a directory '../dataset' for the above few-shot segmentation datasets and appropriately place each dataset to have following directory structure:

../                         # parent directory
├── ./                      # current (project) directory
│   ├── common/             # (dir.) helper functions
│   ├── data/               # (dir.) dataloaders and splits for each FSSS dataset
│   ├── model/              # (dir.) implementation of VRP-SAM 
│   ├── segment-anything/   # code for SAM
│   ├── README.md           # intstruction for reproduction
│   ├── train.py            # code for training HSNet
│   └── SAM2Pred.py         # code for prediction module
│    
├── resnet50_v2.pth
├── vgg16.pth
│    
└── dataset/
    ├── VOC2012/            # PASCAL VOC2012 devkit
    │   ├── Annotations/
    │   ├── ImageSets/
    │   ├── ...
    │   └── SegmentationClassAug/
    └── COCO2014/           
        ├── annotations/
        │   ├── train2014/  # (dir.) training masks (from Google Drive) 
        │   ├── val2014/    # (dir.) validation masks (from Google Drive)
        │   └── ..some json files..
        ├── train2014/
        └── val2014/

Training

sh scripts/train_pascal.sh  
sh scripts/train_coco.sh  

BibTeX

If you use this code for your research, please consider citing:

@article{park2025foreground,
  title={Foreground-Covering Prototype Generation and Matching for SAM-Aided Few-Shot Segmentation},
  author={Park, Suho and Lee, SuBeen and Seong, Hyun Seok and Yoo, Jaejoon and Heo, Jae-Pil},
  journal={arXiv preprint arXiv:2501.00752},
  year={2025}
}