FedVSR: Towards Model-Agnostic Federated Learning in Video Super-Resolution
Open Access
- 4 April 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 118-129
Abstract
Video super-resolution (VSR) aims to enhance low-resolution videos by leveraging both spatial and temporal information. While deep learning has led to impressive progress, it typically requires centralized data, which raises privacy concerns. Federated learning (FL) offers a privacy-friendly solution, but general FL frameworks often struggle with low-level vision tasks, resulting in blurry, low-quality outputs. To address this, we introduce FedVSR, the first FL framework specifically designed for VSR. It is architecture-agnostic and stateless, and introduces a lightweight loss function based on the Discrete Wavelet Transform (DWT) to better preserve high-frequency details during local training. Additionally, a loss-aware aggregation strategy combines both DWT-based and task-specific losses to guide global updates effectively. Extensive experiments across multiple VSR models and datasets show that FedVSR not only improves perceptual video quality (up to +0.89 dB PSNR, +0.0370 SSIM, -0.0347 LPIPS and 4.98 VMAF) but also achieves these gains with close to zero computation and communication overhead compared to its rivals. These results demonstrate Fed-VSR's potential to bridge the gap between privacy, efficiency, and perceptual quality, setting a new benchmark for federated learning in low-level vision tasks. Please refer to this link for the code. https://github.com/alimd94/FedVSRKeywords
All Related Versions
This publication has 40 references indexed in Scilit:
- A comparative analysis of federated learning strategies for short-term photovoltaic power forecastingEnergy Reports, 2025
- Navigating Privacy: A Global Comparative Analysis of Data Protection LawsIET Information Security, 2025
- FedSR: Federated Learning for Image Super-Resolution via detail-assisted contrastive learningKnowledge-Based Systems, 2024
- Real-world video superresolution enhancement method based on the adaptive down-sampling modelScientific Reports, 2024
- A Survey for Federated Learning Evaluations: Goals and MeasuresIEEE Transactions on Knowledge and Data Engineering, 2024
- Model aggregation techniques in federated learning: A comprehensive surveyFuture Generation Computer Systems, 2023
- Video super-resolution based on deep learning: a comprehensive surveyArtificial Intelligence Review, 2022
- Federated Learning: A Survey on Enabling Technologies, Protocols, and ApplicationsIEEE Access, 2020
- Fractional-Grey Wolf optimizer-based kernel weighted regression model for multi-view face video super resolutionInternational Journal of Machine Learning and Cybernetics, 2017
- A wavelet transform based contrast enhancement method for underwater acoustic imagesMultidimensional Systems and Signal Processing, 2017