Home

Awesome

Onion-Peel Networks for Deep Video Completion

Seoung Wug Oh, Sungho Lee, Joon-Young Lee, Seon Joo Kim

ICCV 2019

[paper]

Onion-Peel Networks for Deep Video Completion

- This repository contains a demo software for OPN with following applications

  1. Reference guided image completion (group photo inpainting)
  2. Video completion

- Requirements

- How to Use

Environment setup

conda create --name opn python=3.6
source activate opn

pip install opencv-contrib-python pillow

conda install pytorch=0.4.0 cuda90 -c pytorch
conda install torchvision

Download weights

Place it the same folder with demo scripts
wget -O OPN.pth "https://www.dropbox.com/s/sxo25p12sfc7na7/OPN.pth?dl=1"
wget -O TCN.pth "https://www.dropbox.com/s/nihciqj551xv7a8/TCN.pth?dl=1"

Run

1) Group photo inpainting
python demo_group_image.py --input 3e91f10205_2
2) Video inpainting
python demo_video.py --input parkour

Test your own images/videos

Prepare your images/videos in Image_inputs/[name] or Video_inputs/[name], in the same format and naming rule with the provided examples.

then, run

python demo_group_image.py --input [name]

or,

python demo_video.py --input [name]

- Reference

If you find our paper and repo useful, please cite our paper. Thanks!

Onion-Peel Networks for Deep Video Completion
Seoung Wug Oh, Sungho Lee, Joon-Young Lee, Seon Joo Kim
ICCV 2019

- Related Project

Please check out our another approach for video inpaining!

Copy-and-Paste Networks for Deep Video Inpainting
Sungho Lee, Seoung Wug Oh, DaeYeun Won,  Seon Joo Kim
ICCV 2019

[paper] [github]

- Datasets

This repository partially contains frames sampled from Youtube-VOS videos for the group photo application, and DAVIS videos and masks from Huang et al. for the video inpainting.

- Terms of Use

This software is for non-commercial use only. The source code is released under the Attribution-NonCommercial-ShareAlike (CC BY-NC-SA) Licence (see this for details)