FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation

Ziyuan Tao, Chuanzhi Xu, Sandaru Jayawardana, Wei Bao, Kanchana Thilakarathna +1 more
12/21/2025
cs.CVcs.AIcs.MM

Official code: zyt-599/FedVideoMAE (5 stars, last commit 2026-01-28).

Abstract

The rapid growth of short-form video platforms increases the need for privacy-preserving moderation, as cloud-based pipelines expose raw videos to privacy risks, high bandwidth costs, and inference latency. To address these challenges, we propose an on-device federated learning framework for video violence detection that integrates self-supervised VideoMAE representations, LoRA-based parameter-efficient adaptation, and defense-in-depth privacy protection. Our approach reduces the trainable parameter count to 5.5M (~3.5% of a 156M backbone) and incorporates DP-SGD with configurable privacy budgets and secure aggregation. Experiments on RWF-2000 with 40 clients achieve 77.25% accuracy without privacy protection and 65-66% under strong differential privacy, while reducing communication cost by $28.3\times$ compared to full-model federated learning. The code is available at: {https://github.com/zyt-599/FedVideoMAE}

View on arXivView PDF

Code Implementations(1)

zyt-599/FedVideoMAEOfficial100%
50PythonOct 30, 20258 months ago

Similar Papers

Federated Learning for Video Violence Detection: Complementary Roles of Lightweight CNNs and Vision-Language Models for Energy-Efficient Use

85% match2025

JOSENet: A Joint Stream Embedding Network for Violence Detection in Surveillance Videos

83% match1 repo82024

Balancing Accuracy and Training Time in Federated Learning for Violence Detection in Surveillance Videos: A Study of Neural Network Architectures

81% match2023

Frugal Federated Learning for Violence Detection: A Comparison of LoRA-Tuned VLMs and Personalized CNNs

79% match2025

FedDiSC: A Computation-efficient Federated Learning Framework for Power Systems Disturbance and Cyber Attack Discrimination

79% match2023

TAPFed: Threshold Secure Aggregation for Privacy-Preserving Federated Learning

79% match2025

Cite this paper

@article{tao2025fedvideomae,
  title  = {FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation},
  author = {Ziyuan Tao and Chuanzhi Xu and Sandaru Jayawardana and Wei Bao and Kanchana Thilakarathna and Teng Joon Lim},
  year   = {2025},
  eprint = {2512.18809},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2512.18809v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2512.18809.svg)](https://codeofpaper.com/paper/2512.18809)

Discussion