3D InPainiting Demo

To try it on your own, follow these steps

• Prepare environment

!pip install -r requirements.txt

!pip install torch==1.4.0+cu100 torchvision==0.5.0+cu100 -f https://download.pytorch.org/whl/torch_stable.html

!sudo apt install sed

• Download script and pretrained model

%cd /content/
!git clone https://github.com/vt-vl-lab/3d-photo-inpainting.git
%cd 3d-photo-inpainting
!sh download.sh

• Switch off off-screen rendering

!sed -i 's/offscreen_rendering: True/offscreen_rendering: False/g' 3d-photo-inpainting/argument.yml

• Please upload .jpg files to /content/3d-photo-inpainting-master/image/
  • You can run this step multiple times to upload multiple .jpg files.

%cd 3d-photo-inpainting/image
from google.colab import files
uploaded = files.upload()
for fn in uploaded.keys():
  print('User uploaded file "{name}" with length {length} bytes'.format(
      name = fn, length = len(uploaded[fn])))
%cd ..

• Execute the 3D Photo Inpainting

!python main.py --config argument.yml

References

@inproceedings{Shih3DP20,
  author = {Shih, Meng-Li and Su, Shih-Yang and Kopf, Johannes and Huang, Jia-Bin},
  title = {3D Photography using Context-aware Layered Depth Inpainting},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2020}
}