GraphFPN: Graph Feature Pyramid Network for Object Detection

Download graph-FPN-main.zip

For training , run:

python train.py

For test with Graph_fpn, run

python test.py

If You need COCO API for test, you can download from here.

Folder structure

${ROOT}
└── checkpoint/
└── COCO/    
│   └── coco/
│   │    ├── .config 
│   │    ├── 2017/
│   │
│   ├── downloads/
│
│
└── data_demo/
|   ├── data/
|   |    ├── coco
|   |    ├── checkpoint
|   ├── data.zip
|
├── results/
├── src/     
|   ├── configs/
|   |    ├── configs.py
|   |
|   ├── detection/
|   |    ├── datasets/
|   |    |      ├── coco.py
|   |    ├── utils/
|   |
|   ├── model/
|   ├── init_path.py
|   ├── demo.py
|   ├── train.py
|   ├── test.py
├── README.md 
└── requirements.txt

References

[1] Graph-FPN: GraphFPN: Graph Feature Pyramid Network for Object Detection

In addition, we provide more detection frameworks that can support GraphFPN

Download graph-mmdet.zip 

this code uses mmdetecion as the base framework, you can set yourself env based on mmdetection
this can simply run

sh train.sh

get the result of Contextual Graph Layers (CGL-1) in graphFPN, however, you should add other components from graph-FPN-main.zip to run the complete GraphFPN. Note that, based on the code of graph-mmdet.zip, you can easily construct the complete graph-fpn strcuture. Please reference the code of graph-FPN-main.zip.

Note that graph-FPN-main.zip is originated from the following link

https://github.com/lhcezx/Graph-FPN.git

Based on the above code and mmdetection, we will improve our the codebase of graphFPN better and make it clearer.

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

@inproceedings{GraphFPN,
  author  = {Gangming Zhao and Weifeng Ge and Yizhou Yu},
  title     = {GraphFPN: Graph Feature Pyramid Network for Object Detection},
  booktitle = {IEEE/CVF International Conference on Computer Vision (ICCV)},
  pages     = {2743--2752},
  publisher = {IEEE},
  year      = {2021},
}