luogaara / Background-Matting

Background Matting: The World is Your Green Screen

Home Page:

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Background Matting: The World is Your Green Screen

alt text

By Soumyadip Sengupta, Vivek Jayaram, Brian Curless, Steve Seitz, and Ira Kemelmacher-Shlizerman

This paper will be presented in IEEE CVPR 2020.

Project Page

Go to Project page for additional details and results.

Paper (Arxiv)

Project members


This work is licensed under the Creative Commons Attribution NonCommercial ShareAlike 4.0 License.

Getting Started

Clone repository:

git clone

Please use Python 3. Create an Anaconda environment and install the dependencies. Our code is tested with Pytorch=1.1.0, Tensorflow=1.14 with cuda10.0

conda create --name back-matting python=3.6
conda activate back-matting

Make sure CUDA 10.0 is your default cuda. If your CUDA 10.0 is installed in /usr/local/cuda-10.0, apply

export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64
export PATH=$PATH:/usr/local/cuda-10.0/bin

Install PyTorch, Tensorflow (needed for segmentation) and dependencies

conda install pytorch=1.1.0 torchvision cudatoolkit=10.0 -c pytorch
pip install tensorflow-gpu=1.14.0
pip install -r requirements.txt

Note: The code is likely to work on other PyTorch and Tensorflow versions compatible with your system CUDA. If you already have a working environment with PyTorch and Tensorflow, only install dependencies with pip install -r requirements.txt. If our code fails due to different versions, then you need to install specific CUDA, PyTorch and Tensorflow versions.

Run the inference code on sample images


To perform Background Matting based green-screening, you need to capture:

  • (a) Image with the subject (use _img.png extension)
  • (b) Image of the background without the subject (use _back.png extension)
  • (c) Target background to insert the subject (place in data/background)

Use sample_data/ folder for testing and prepare your own data based on that.

Pre-trained model

Please download the pre-trained models from Google Drive and place Models/ folder inside Background-Matting/.


  1. Segmentation

Background Matting needs a segmentation mask for the subject. We use tensorflow version of Deeplabv3+.

cd Background-Matting/
git clone
cd models/research/
export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim
cd ../..
python -i sample_data/input

You can replace Deeplabv3+ with any segmentation network of your choice. Save the segmentation results with extension _masksDL.png.

  1. Alignment
  • For hand-held camera, we need to align the background with the input image as a part of pre-processing. We apply simple hoomography based alignment.
  • We ask users to disable the auto-focus and auto-exposure of the camera while capturing the pair of images. This can be easily done in iPhone cameras (tap and hold for a while).

Run python -i sample_data/input for pre-processing. It aligns the background image _back.png and changes its bias-gain to match the input image _img.png

Background Matting

python -m real-hand-held -i sample_data/input/ -o sample_data/output/ -tb sample_data/background/0001.png

For images taken with fixed camera (with a tripod), choose -m real-fixed-cam for best results. -m syn-comp-adobe lets you use the model trained on synthetic-composite Adobe dataset, without real data (worse performance).

Notes on capturing images

For best results capture images following these guidelines:

  • Choose a background that is mostly static, can be both indoor and outdoor.
  • Avoid casting any shadows of the subject on the background.
    • place the subject atleast few feets away from the background.
    • if possible adjust the lighting to avoid strong shadows on the background.
  • Avoid large color coincidences between subject and background. (e.g. Do not wear a white shirt in front of a white wall background.)
  • Lock AE/AF (Auto-exposure and Auto-focus) of the camera.
  • For hand-held capture, you need to:
    • allow only small camera motion by continuing to holding the camera as the subject exists the scene.
    • avoid backgrounds that has two perpendicular planes (homography based alignment will fail) or use a background very far away.
    • The above restirctions do not apply for images captured with fixed camera (on a tripod)

Training code

We will also release the training code, which will allow users to train on labelled data and also on unlabelled real data.

Coming soon ...


We collected 50 videos with both fixed and hand-held camera in indoor and outdoor settings. We plan to release this data to encourage future research on improving background matting.

Coming soon ...


We are eager to hear how our algorithm works on your images/videos. If the algorithm fails on your data, please feel free to share it with us at This will help us in improving our algorithm for future research. Also, feel free to share any cool results.


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

  title={Background Matting: The World is Your Green Screen},
  author = {Soumyadip Sengupta and Vivek Jayaram and Brian Curless and Steve Seitz and Ira Kemelmacher-Shlizerman},
  booktitle={Computer Vision and Pattern Regognition (CVPR)},


Background Matting: The World is Your Green Screen


Language:Python 100.0%