Home

Awesome

Event-Based Motion Magnification

Project Page | Video | Paper | Data <br>

ECCV 2024

Yutian Chen, Shi Guo, Fangzheng Yu, Feng Zhang, Jinwei Gu, Tianfan Xue <br><br>

<p align="left" width="100%"> <img src="docs/static/images/teaser.gif" width="90%" > </p> <!-- - [ ] Release the real-captured dataset - [ ] Release the synthetic testset. - [x] Release the training and testing code. - [x] Release the pretrained model. -->

Installation

This implementation based on BasicSR which is a open source toolbox for image/video restoration tasks.

git clone https://github.com/OpenImagingLab/emm.git
cd emm
pip install -r requirements.txt
python setup.py develop --no_cuda_ext

Running code

Dataset

Refer to the REAME.md for instructions on preparing training data. We also provide a mini batch of train set and real-captured test set on Google Drive as an example.

Train

To train the EMM model:

# Modify the dataroot options/train.yml 
bash train.sh

Test

To test the EMM model with real-captured video:

# Modify the dataroot and the temporal filter parameters in options/train.yml 
bash test.sh

Citations

@article{chen2024eventbased,
      title={Event-Based Motion Magnification}, 
      author={Yutian Chen and Shi Guo and Fangzheng Yu and Feng Zhang and Jinwei Gu and Tianfan Xue},
      journal={arXiv preprint arXiv:2402.11957},
      year={2024}
}