EvTexture (ICML 2024)

Official Pytorch implementation for the "EvTexture: Event-driven Texture Enhancement for Video Super-Resolution" paper (ICML 2024).

🌐 Project | 📃 Paper | 🖼️ Poster

Authors: Dachun Kai^📧️, Jiayao Lu, Yueyi Zhang^📧️, Xiaoyan Sun, University of Science and Technology of China

Feel free to ask questions. If our work helps, please don't hesitate to give us a ⭐!

🚀 News

• Release training code
• 2024/06/28: Release details to prepare datasets
• 2024/06/08: Publish docker image
• 2024/06/08: Release pretrained models and test sets for quick testing
• 2024/06/07: Video demos released
• 2024/05/25: Initialize the repository
• 2024/05/02: 🎉 🎉 Our paper was accepted in ICML'2024

🔖 Table of Content

1. Video Demos
2. Code
3. Citation
4. Contact
5. License and Acknowledgement

🔥 Video Demos

A $4\times$ upsampling results on the Vid4 and REDS4 test sets.

Vid4_City.mp4

Vid4_Foliage.mp4

REDS_000.mp4

REDS_011.mp4

Code

Installation

• Dependencies: Miniconda, CUDA Toolkit 11.1.1, torch 1.10.2+cu111, and torchvision 0.11.3+cu111.

• Run in Conda

  conda create -y -n evtexture python=3.7
  conda activate evtexture
  pip install torch-1.10.2+cu111-cp37-cp37m-linux_x86_64.whl
  pip install torchvision-0.11.3+cu111-cp37-cp37m-linux_x86_64.whl
  git clone https://github.com/DachunKai/EvTexture.git
  cd EvTexture && pip install -r requirements.txt && python setup.py develop

• Run in Docker 👏

  Note: before running the Docker image, make sure to install nvidia-docker by following the official instructions.

  [Option 1] Directly pull the published Docker image we have provided from Alibaba Cloud.

  docker pull registry.cn-hangzhou.aliyuncs.com/dachunkai/evtexture:latest

  [Option 2] We also provide a Dockerfile that you can use to build the image yourself.

  cd EvTexture && docker build -t evtexture ./docker

  The pulled or self-built Docker image containes a complete conda environment named evtexture. After running the image, you can mount your data and operate within this environment.

  source activate evtexture && cd EvTexture && python setup.py develop

Test

1. Download the pretrained models from (Releases / Onedrive / Google Drive / Baidu Cloud(n8hg)) and place them to experiments/pretrained_models/EvTexture/. The network architecture code is in evtexture_arch.py.

   • EvTexture_REDS_BIx4.pth: trained on REDS dataset with BI degradation for $4\times$ SR scale.
   • EvTexture_Vimeo90K_BIx4.pth: trained on Vimeo-90K dataset with BI degradation for $4\times$ SR scale.

2. Download the preprocessed test sets (including events) for REDS4 and Vid4 from (Releases / Onedrive / Google Drive / Baidu Cloud(n8hg)), and place them to datasets/.

   • Vid4_h5: HDF5 files containing preprocessed test datasets for Vid4.

   • REDS4_h5: HDF5 files containing preprocessed test datasets for REDS4.

3. Run the following command:

   • Test on Vid4 for 4x VSR:

     ./scripts/dist_test.sh [num_gpus] options/test/EvTexture/test_EvTexture_Vid4_BIx4.yml
     

   • Test on REDS4 for 4x VSR:

     ./scripts/dist_test.sh [num_gpus] options/test/EvTexture/test_EvTexture_REDS4_BIx4.yml
     

     This will generate the inference results in results/. The output results on REDS4 and Vid4 can be downloaded from (Releases / Onedrive / Google Drive / Baidu Cloud(n8hg)).

Data Preparation

• Both video and event data are required as input, as shown in the snippet. We package each video and its event data into an HDF5 file.

• Example: The structure of calendar.h5 file from the Vid4 dataset is shown below.

  calendar.h5
  ├── images
  │   ├── 000000 # frame, ndarray, [H, W, C]
  │   ├── ...
  ├── voxels_f
  │   ├── 000000 # forward event voxel, ndarray, [Bins, H, W]
  │   ├── ...
  ├── voxels_b
  │   ├── 000000 # backward event voxel, ndarray, [Bins, H, W]
  │   ├── ...
  

• To simulate and generate the event voxels, refer to the dataset preparation details in DataPreparation.md.

Inference on your own video

❤️ Seeking Collaboration: For issues #6 and #7, our method can indeed perform inference on videos without event data. The solution is to use an event camera simulator, such as vid2e, to generate event data from the video, and then input both the video data and the generated event data into our model. This part, however, may require extensive engineering work to package everything into a script, as detailed in DataPreparation.md. We currently do not have enough time to undertake this task, so we are looking for collaborators to join us in this effort! 😊

😊 Citation

If you find the code and pre-trained models useful for your research, please consider citing our paper. 😃

@inproceedings{kai2024evtexture,
  title={Ev{T}exture: {E}vent-driven {T}exture {E}nhancement for {V}ideo {S}uper-{R}esolution},
  author={Kai, Dachun and Lu, Jiayao and Zhang, Yueyi and Sun, Xiaoyan},
  booktitle={International Conference on Machine Learning},
  year={2024},
  organization={PMLR}
}

Contact

If you meet any problems, please describe them in issues or contact:

• Dachun Kai: [email protected]

License and Acknowledgement

This project is released under the Apache-2.0 license. Our work is built upon BasicSR, which is an open source toolbox for image/video restoration tasks. Thanks to the inspirations and codes from RAFT and event_utils.