VIDNet

Code for the VIDNet (BMVC 2021)

Installation

The base code is from RVOS.

• Install requirements pip install -r requirements.txt

Data

DAVIS 2016

Download the DAVIS 2016 dataset from their website at 480p resolution. Create a folder named databasesin the parent folder of the root directory of this project and put there the database in a folder named DAVIS2016. The root directory (VIDNetfolder) and the databases folder should be in the same directory.

Inpainted DAVIS 2016

Follow the instruction on VINet to inpaint DAVIS and obtain VI inpainting result: https://github.com/mcahny/Deep-Video-Inpainting

Follow the instruction on OPNet to inpaint DAVIS and obtain OP inpainting result: https://github.com/seoungwugoh/opn-demo

Follow the instruction on CPNet to inpaint DAVIS and obtain CP inpainting result: https://github.com/shleecs/Copy-and-Paste-Networks-for-Deep-Video-Inpainting

Inpainted DAVIS 2016 ELA frame extraction

Run ela.py and change the path of davis_train and davis_test to corresponding inpainting DAVIS folder.

Training

• Train the model for video inpainting detection with python train_vi.py -model_name model_name. Checkpoints and logs will be saved under ../models/model_name.

• Other arguments can be passed as well. For convenience, scripts to train with typical parameters are provided under scripts/. Simply run the the following:

1. cd src/misc
2. change __C.PATH.SEQUENCES and __C.PATH.SEQUENCES2 in config.py to be the training data path (e.g., VI, OP, CP inpainting path)
3. cd ../../script
4. update model_name for model storage
5. run train_davis.sh

Evaluation

We provide bash scripts to evaluate models for the DAVIS 2016 and FVI datasets. You can find them under the scripts folder.

For DAVIS evaluation:

1. cd src/misc
2. change __C.PATH.SEQUENCES in config.py to be the testing data path (e.g., VI, OP, CP)
3. cd ../../scripts
4. update model_name in eval_davis.sh
5. run eval_davis.sh For FVI evaluation:
6. cd scripts
7. update model_name in eval_fvi.sh
8. run eval_fvi.sh

Citation

If this code or dataset helps your research, please cite our paper:

@inproceedings{zhou2021vid,
  title={Deep Video Inpainting Detection},
  author={Zhou, Peng and Yu, Ning and Wu, Zuxuan and Davis, Larry S and Shrivastava, Abhinav and Lim, Ser Nam},
  booktitle = {BMVC},
  year={2021}
}