The official code of "SILCO: Show a Few Images, Localize the Common Object"(ICCV2019)

Introduction

Starting from a few weakly-supervised support images and a query image, we are able to localize the common object in the query image without the need for a single box annotation. We only know that there is a common object, we do not know where and what the common object is. For more details, please refer to our paper.

For more details of our framework, please refer to our paper or project website.

Acknowledgement

The code is based on https://github.com/amdegroot/ssd.pytorch,https://github.com/vgsatorras/few-shot-gnn,https://github.com/AlexHex7/Non-local_pytorch, the mAP evaluation is based on https://github.com/rafaelpadilla/Object-Detection-Metrics, thanks for their contribution for open sourcing.

Instruction

• set up your environment by anaconda,(python3.5, torch 0.3.1)
• pip install cython
• pip install tqdm
• pip install termcolor
• pip install tabulate
• pip install matplotlib
• pip install --no-cache --upgrade git+https://github.com/dongzhuoyao/pytorchgo.git
• pip install opencv-python
• pip install image
• pip install torchvision==0.2
• install pycocotools via https://github.com/cocodataset/cocoapi/tree/master/PythonAPI
• download vgg16_reducedfc.pth, put it in root path.
• set dataset path of Pasval VOC and COCO14.

ln -s your_local_coco14_path coco14
ln -s your_local_pascalvoc_path VOCdevkit

Training

CUDA_VISIBLE_DEVICES=0 python silco.py  --dataset cl_voc07 --subset 0

Testing

CUDA_VISIBLE_DEVICES=0 python silco.py  --dataset cl_voc07 --subset 0 --test

Our pretrained weight can be found https://drive.google.com/open?id=1zNIHchYkIo1hg64usgonAJtK9vAQcQr-

Dataset

Check data/list directory.

Citation

If you find the code or trained models useful, please consider citing:

@inproceedings{hu2019silco,
  title={{SILCO}: show a few images, localize the common object},
  author={Tao HU and Pascal Mettes and Jia-Hong Huang and Cees G. M. Snoek},
  booktitle={ICCV},
  year={2019}
}

Question

For question, please send me email: t+a+o+h+u+6+2+0 at gmail.com(remove '+').