bmbc's Introduction

BMBC

Junheum Park, Keunsoo Ko, Chul Lee, and Chang-Su Kim

Official PyTorch Code for "BMBC: Bilateral Motion Estimation with Bilateral Cost Volume for Video Interpolation" [paper]

Requirements

PyTorch 1.3.1 (Other versions can cause different results)
CUDA 10.0
CuDNN 7.6.5
python 3.6

Installation

Create conda environment:

    $ conda create -n BMBC python=3.6 anaconda
    $ conda activate BMBC
    $ pip install opencv-python
    $ conda install pytorch==1.3.1 torchvision cudatoolkit=10.0 -c pytorch

Download repository:

    $ git clone https://github.com/JunHeum/BMBC.git

Download pre-trained model parameters:

    $ unzip BMBC_Weights.zip

Usage

Generate an intermediate frame at t=0.5 on your pair of frames:

    $ python run.py --first images/im1.png --second images/im3.png --output images/im2.png

Generate an intermediate frame at arbitrary time t:

    $ python run.py --first images/im1.png --second images/im3.png --output images/im2_025.png --time_step 0.25

Citation

Please cite the following paper if you feel this repository useful.

    @inproceedings{BMBC,
        author    = {Junheum Park, Keunsoo Ko, Chul Lee,and Chang-Su Kim}, 
        title     = {BMBC: Bilateral Motion Estimation with Bilateral Cost Volume for Video Interpolation}, 
        booktitle = {European Conference on Computer Vision},
        year      = {2020}
    }