Self-supervised multi-scale pyramid fusion networks for realistic bokeh effect rendering

| Paper |

If you are interested in my work, you can contact me by email(zhifengwang686@gmail.com), I will check the email regularly. thanks!

1、Dataset

You could get the EBB! dataset by registering here.and put it in the train folder.

Train split: data/train.csv

Test split (val294 set): data/test.csv

2、Installation

git clone https://github.com/zfw-cv/MPFNet.git
cd MPFNet
pip install -r requirements.txt

3、Train

You can download the pre-training model by here .And put it in the checkpoints folder.

python train.py

4、Test

python test.py

5、Val

To test our effects more easily, you can directly use the results obtained from our pre-trained weights file.

python val.py

6、Citation

If you find our work useful in your research, please cite our paper.

@article{wang2022self,
  title={Self-supervised multi-scale pyramid fusion networks for realistic bokeh effect rendering},
  author={Wang, Zhifeng and Jiang, Aiwen and Zhang, Chunjie and Li, Hanxi and Liu, Bo},
  journal={Journal of Visual Communication and Image Representation},
  pages={103580},
  year={2022},
  publisher={Elsevier}
}