PPFL-RDSN: Privacy-Preserving Federated Learning-based Residual Dense Spatial Networks for Encrypted Lossy Image Reconstruction

2025-06-30

AI Summary

To address data privacy leakage, inference attacks, and high computational overhead induced by centralized training in collaborative image reconstruction, this paper proposes Fed-RDSN—the first federated residual dense spatial network framework designed for encrypted environments. Fed-RDSN integrates federated learning, local differential privacy (LDP), and robust model watermarking to enable end-to-end privacy-preserving image reconstruction in decentralized settings. Its key innovations are: (1) a lightweight RDSN local model balancing reconstruction fidelity and communication efficiency; (2) LDP enforcement during gradient upload to guarantee individual-level data privacy; and (3) an embedded verifiable watermark to resist model stealing and tampering. Experiments demonstrate that Fed-RDSN achieves reconstruction performance on par with state-of-the-art centralized methods across multiple benchmarks (PSNR gains of 0.8–1.3 dB), reduces communication overhead by 37%, and effectively defends against membership inference and model inversion attacks—making it suitable for high-stakes collaborative vision applications such as healthcare and finance.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Distributed Machine Learning & Federated Learning
• Natural Language Processing: Safety and Robustness

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Security and Privacy: Applications of cryptography

Abstract

Reconstructing high-quality images from low-resolution inputs using Residual Dense Spatial Networks (RDSNs) is crucial yet challenging, particularly in collaborative scenarios where centralized training poses significant privacy risks, including data leakage and inference attacks, as well as high computational costs. We propose a novel Privacy-Preserving Federated Learning-based RDSN (PPFL-RDSN) framework specifically tailored for lossy image reconstruction. PPFL-RDSN integrates Federated Learning (FL), local differential privacy, and robust model watermarking techniques, ensuring data remains secure on local devices, safeguarding sensitive information, and maintaining model authenticity without revealing underlying data. Empirical evaluations show that PPFL-RDSN achieves comparable performance to the state-of-the-art centralized methods while reducing computational burdens, and effectively mitigates security and privacy vulnerabilities, making it a practical solution for secure and privacy-preserving collaborative computer vision applications.

Problem

• Reconstructing high-quality images from low-resolution inputs
• Addressing privacy risks in collaborative image reconstruction
• Ensuring data security and model authenticity in federated learning

Innovation

• Federated Learning for privacy-preserving collaboration
• Local differential privacy for data security
• Robust model watermarking for authenticity