FedVSR: Towards Model-Agnostic Federated Learning in Video Super-Resolution (Under Review)

This is the official repository for FedVSR which contains code and experiments, a method to train Video Super-Resolution (VSR) models under the Federated Learning (FL) paradigm. The goal is to enable high-quality VSR while ensuring data privacy by not centralizing raw video data.

Key Contributions

✅ Introduced a model-agnostic and stateless FL framework for VSR

✅ Developed a Discrete Wavelet Transform-based loss function to preserve high-frequency details and enhance reconstruction quality.

✅ Proposed a loss-aware weighted aggregation method with a decay strategy.

✅ Achieved superior PSNR and SSIM across multiple VSR models and datasets.

✅ First framework addressing Federated Learning for VSR.

Project Overview

• What is FedVSR?
  FedVSR is a federated training framework that allows multiple clients to train a shared video super-resolution model without sharing their raw data. This approach preserves data privacy and is suitable for scenarios where raw video footage is sensitive.

Note: This code is released under review status; therefore, no direct citations are referenced in this repo.

---

Repository Structure

.
├── Kinetics_Scripts/
│   └── copy_movie.py          # Script for extracting & copying related movies
│
├── VRTandRVRT/
│   ├── vrt_and_rvrt_{ALG}.py  # Scripts to run all trainings with different methods.     
│   ├── main_test_{MODEL}      # Scripts to run tests on different models    
│   └── 
│
├── IART/
│   ├── iart_{ALG}.py          # Scripts to run all trainings with different methods.        
│   └── test_scripts/test_*.py # Scripts to run tests  
│
├── requirements.txt      # Python package requirements
└── README.md             # Main documentation

---

Installation & Setup

1. Clone the repository

     git clone https://github.com/alimd94/FedVSR.git
     cd FedVSR
   

2. Set up Python environment

    python3 -m venv fedvsr_env
    source fedvsr_env/bin/activate  
    pip install --upgrade pip
    pip install -r requirements.txt
   
   
   

Prepare Data:

To prepare the dataset, follow BasicSR. After completing the preparation, the directory structure should be as follows:

datasets/
├──REDS/
│   └──val_REDS4_sharp
│   └──val_REDS4_sharp_bicubic

Training

For VRT and RVRT prepare the related json in the file and the run the scripts:

# VSR trained on FedAvg
python VRTandRVRT/vrt_and_rvrt_fedavg.py
# VSR trained on FedVSR
python VRTandRVRT/vrt_and_rvrt_fedvsr.py

For IART:

# VSR trained on FedAvg
python IART/iart_fedavg.py
# VSR trained on FedVSR
python IART/iart_fedvsr.py

Testing

You can test using 2 different datasets VID4 and REDS:

python VRTandRVRT/main_test_rvrt.py 

python VRTandRVRT/main_test_vrt.py 

To test IART you would use:

python IART/test_scripts/test_IART_REDS4_N6.py 
python IART/test_scripts/test_IART_Vid4_N6.py

Results

Code References

We use the official implementations of the following models:

• VRT and RVRT:
  The code for training and evaluating VRT and RVRT models is based on the official KAIR Repository

• IART:
  The implementation of IART is based on the official code released by IART Repository.

• Federated Learning Framework:
  We leverage the Flower framework for implementing and managing federated learning workflows.

Please refer to the respective repositories for additional details on model architecture, training strategies, and original paper references.

Citation

This work is currently under review. A formal citation will be added once the review process is complete. In the meantime, feel free to ⭐ star the repository if you find it helpful! 😊